#data-science-and-ml | Python | Page 274

olive lichen Dec 8, 2020, 2:29 AM

#

and the most fucking annoying thing is it was working like ten minutes ago

#

and nothing changed

fiery fossil Dec 8, 2020, 3:46 AM

#

works fine on my side..are you sure liberalMindZScores is a list only?

austere swift Dec 8, 2020, 5:22 AM

#

do you happen to have a conflicting round() function?

lapis sequoia Dec 8, 2020, 5:58 AM

#

oops

azure stump Dec 8, 2020, 7:40 AM

#

https://asr373.medium.com/who-is-he-3f6ef4f9e244

Medium

Who is he?

Can he help Data Scientists?

ocean pumice Dec 8, 2020, 8:56 AM

#

this stuff is insane https://alien-project.org/ i let it run all night long 🙂

artificial life enviroment

ocean pumice Dec 8, 2020, 9:24 AM

#

damn... i need to downgrade to cuda 10.1 to use tensorflow~~ ?~~

ocean pumice Dec 8, 2020, 9:52 AM

#

nvm, it's not a downgrade, both can be installed independently

soft dock Dec 8, 2020, 12:36 PM

#

I don't know how many people know about this, but you guys should try checking out Cython with JAX. Training has never been quicker.

worthy scarab Dec 8, 2020, 2:08 PM

#

Does anybody know how to do a chi square distribution on python? I have my chi square value and data points with a curve of best fit in the form of an exponential but dont know how to make the pdf

lapis sequoia Dec 8, 2020, 2:12 PM

#

worthy scarab Does anybody know how to do a chi square distribution on python? I have my chi s...

is scipy an option? https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.chi2.html

#

anybody know why log scale ticks in matplotlib only give 8 ticks between each power of ten? shouldn't it be 9? how should i read these values otherwise?

📎 unknown.png

normal sinew Dec 8, 2020, 2:53 PM

#

hey

#

who can help me with some practice questions for mathematics for computer science

kindred valley Dec 8, 2020, 2:54 PM

#

yes i need some help with some questions too

split eagle Dec 8, 2020, 5:00 PM

#

I am working in Jupyter notebook and trying to filter a dataframe of medical studies to only include those focused on cancer. I have created a list of 31 cancer-related strings (which I made into a list I called cancer_mesh) and want to filter the df to only include studies for which at least one of the strings in this list appear in the column "mesh_terms". Here is the problem: when I used .isin(cancer_mesh), the output is too restrictive because only studies where the string in "mesh_terms" match the strings in "cancer_mesh" exactly are included. When I try to use .str.contains(cancer_mesh), I get TypeError: unhashable type: 'list'. If I try to use .str.contains() with all the cancer strings inside the parentheses I get TypeError: contains() takes from 2 to 6 positional arguments but 31 were given. I have searched multiple sources online, but haven't found a solution that has worked. Any ideas what I could do to filter this df to include studies where the smaller strings in cancer_mesh appear in the larger strings in the "mesh_terms" columns? I can provide the full list of cancer_mesh strings if you'd like.

worthy scarab Dec 8, 2020, 5:17 PM

#

When i have an xarray and yarray that plots a curve, how can i find the area under that curve?

split eagle Dec 8, 2020, 5:27 PM

#

split eagle I am working in Jupyter notebook and trying to filter a dataframe of medical stu...

FYI: I resolved this problem with the following solution: pattern = '|'.join([f'(?i){x}' for x in cancer_mesh])
df_browse_conditions ['cancer_trial'] = np.where(df_browse_conditions.downcase_mesh_term.str.contains(pattern, na=False),1,0) I will drop all rows where "cancer_mesh" = 0

livid quartz Dec 8, 2020, 6:58 PM

#

How do you get the shap values for Adaboost?

crisp crypt Dec 8, 2020, 6:58 PM

#

@worthy scarab integration

rustic dew Dec 8, 2020, 7:04 PM

#

livid quartz How do you get the shap values for Adaboost?

https://github.com/slundberg/shap works with scikit-learn and you have adaboost in sklearn I believe

GitHub

slundberg/shap

A game theoretic approach to explain the output of any machine learning model. - slundberg/shap

livid quartz Dec 8, 2020, 7:05 PM

#

I've tried that package but I get an error when Applying it to my adamodel

#

But it works for a standard decision tree model

tiny rain Dec 8, 2020, 7:39 PM

#

In a hackathon, how can I implement some Predictive analysis on my data fast ..?

#

Google colab ..?

brazen owl Dec 8, 2020, 8:35 PM

#

Hi everyone

outer fulcrum Dec 8, 2020, 8:41 PM

#

Hey guys

twilit pilot Dec 8, 2020, 8:57 PM

#

Does anyone know an API or anything that will give me the support and resistance lines for a stock if I give it the candlestick data?

torpid cave Dec 8, 2020, 10:22 PM

#

H guys

#

Do you know any easy way of handling missing index information?

#

e,g,

#

info[3] when index 3 does not exists

lilac raven Dec 8, 2020, 10:32 PM

#

Hello, in relation to :
fig, ax = plt.subplots(2, figsize=(x,y)) fig.canvas.manager.window.move(x,y)
. Is there a command that would let me see the current fig size and screen position if I want to manually move it after starting the code?

torpid cave Dec 8, 2020, 10:39 PM

#

@twilit pilot if you have the data, then you can create the graphs yourself

#

Good example: https://plotly.com/python/candlestick-charts/

Candlestick Charts

How to make interactive candlestick charts in Python with Plotly. Six examples of candlestick charts with Pandas, time series, and yahoo finance data.

twilit pilot Dec 8, 2020, 10:41 PM

#

torpid cave <@!747458337511047179> if you have the data, then you can create the graphs your...

i did but after i also want to get the support and resistance lines of a stock

torpid cave Dec 8, 2020, 10:41 PM

#

What are the formulas behind support and resistance/

#

?

#

You could add them manually

frosty flare Dec 8, 2020, 10:44 PM

#

hey guys , its about " OPENCV "

twilit pilot Dec 8, 2020, 10:46 PM

#

torpid cave What are the formulas behind support and resistance/

idk i might have to create my own logic if there is nothing out there

frosty flare Dec 8, 2020, 10:46 PM

#

so i wrote the code to detect and extract faces into new windows , and made it save the detected face when i press i button

#

my problem is if theres multiple faces it does only save the last one which has been detected

torpid cave Dec 8, 2020, 10:47 PM

#

@twilit pilot I am thinking that they are set-up randomly

#

Just saw a quick explanation on them

twilit pilot Dec 8, 2020, 10:49 PM

#

torpid cave <@!747458337511047179> I am thinking that they are set-up randomly

yea because not all stocks are the same, so there is not real formula, but there is some kind of logic behind it

torpid cave Dec 8, 2020, 10:50 PM

#

I wonder how do stock trading sofware plot them

#

I just saw one which used 15-day MA

twilit pilot Dec 8, 2020, 10:51 PM

#

yea im pretty sure fidelity has their own support and resistance algorithm, which i have used before

#

i am trying to create a stock trader, so i want to know how to find the support and resistance

south brook Dec 8, 2020, 10:51 PM

#

We are 3 girls meet every week and exploring the Kaggle titanic competition.
We relay on this notebook : https://www.kaggle.com/arg0n007/titanic-80-accuracy-top-14-random-forest
We need a little mentoring help because we got stuck in Confusion Matrix and Kfold Validation phase.
All we need from you(Mentor) is one hour only ,please who ever can come up forward to help us we will be very happy , thankful and appropriated to him\her.

twilit pilot Dec 8, 2020, 10:52 PM

#

well ik how to find support and resistance, but i dont know how to code it lol

torpid cave Dec 8, 2020, 10:52 PM

#

Hahaha

#

Like visually or with a formula?

twilit pilot Dec 8, 2020, 10:53 PM

#

well step one is to find the turning points of a graph - i did that

torpid cave Dec 8, 2020, 10:53 PM

#

Ok delta

twilit pilot Dec 8, 2020, 10:53 PM

#

step two is to filter those points to get only the relevant points

#

that pretty much it lol, but idk how accurate this will be thats why if there is an api already, i would rather use that

torpid cave Dec 8, 2020, 10:55 PM

#

For financial stuff, things tend to cost money.. just FYI

#

lots of money

twilit pilot Dec 8, 2020, 10:58 PM

#

haha yes ik. perhaps i will have to create my own logic

torpid cave Dec 8, 2020, 10:58 PM

#

If you are interested, there is this nice article in Medium:
https://towardsdatascience.com/programmatic-identification-of-support-resistance-trend-lines-with-python-d797a4a90530

Medium

Programmatic Identification of Support/Resistance Trend lines with ...

Automated trend line analysis using Python done comprehensively and properly using different methods and tools.

twilit pilot Dec 8, 2020, 11:02 PM

#

thanks for this article

twilit brook Dec 8, 2020, 11:44 PM

#

Hey guys

lapis sequoia Dec 8, 2020, 11:44 PM

#

is it possible to write a very very simple SIR disease model in vanilla python

twilit brook Dec 8, 2020, 11:44 PM

#

Are there any professional data scientists on this channel? Wanted to ask a few career questions. If I can steal some of your time pls PM me 🙂

fading wigeon Dec 9, 2020, 12:31 AM

#

lapis sequoia is it possible to write a very very simple SIR disease model in vanilla python

Yes

#

I've written one in Matlab a while ago and I don't think I used anything that wasn't also available in python

#

Also, does anyone have any interest in data science in how it relates to neuroscience? I'm going to an online conference on the subject in 30 minutes, can send an invite if anyone is interested

lapis sequoia Dec 9, 2020, 12:33 AM

#

could I see it maybe?

#

just to get an idea

fading wigeon Dec 9, 2020, 1:02 AM

#

Sorry, it's lost to the sands of time, I'm afraid.

twilit pilot Dec 9, 2020, 1:02 AM

#

bro @fading wigeon ur staus tripped me off 🤣

fading wigeon Dec 9, 2020, 1:02 AM

#

🙂

hollow gull Dec 9, 2020, 2:30 AM

#

@twilit brook ask your questions in the chat and people will respond to them if they can. I believe the community prefers not to use PMs.

lapis sequoia Dec 9, 2020, 2:36 AM

#

guys i have a problem

#

i was using paidml to use gpu from amd, but i think it wasnt working cuz
INFO:plaidml:Opening device "opencl_amd_ellesmere.0"
i think this is cpu still
I had this line on my program os.environ["KERAS_BACKEND"] = "plaidml.keras.backend"
but if i remove it
i get this AttributeError: module 'tensorflow' has no attribute 'get_default_graph'
why?

austere swift Dec 9, 2020, 3:32 AM

#

lapis sequoia i was using paidml to use gpu from amd, but i think it wasnt working cuz ``INFO:...

did you do the plaidml-setup that it says to do in the quickstart?

lapis sequoia Dec 9, 2020, 3:33 AM

#

yes, i did, and it was working

#

if i set the environ it works

#

if i remove that line, i get the error

#

also the setup only shows this

#

https://gyazo.com/bc3df4d2f96b0b0b437adcbb297fa6e6

Gyazo

#

and i think both are cpu

#

not sure

austere swift Dec 9, 2020, 3:43 AM

#

llvm is cpu

#

so the other one is probably gpu

#

unless you have some other weird device or sum idk

lapis sequoia Dec 9, 2020, 3:45 AM

#

still, thats not the point

#

why if i remove the os.environ everything stops working?

austere swift Dec 9, 2020, 3:47 AM

#

can you show your code?

lapis sequoia Dec 9, 2020, 3:47 AM

#

import os
os.environ["KERAS_BACKEND"] = "plaidml.keras.backend"

austere swift Dec 9, 2020, 3:47 AM

#

the os.environ is what's telling keras to use plaidml

#

as to why it shows that error I don't know

#

unless you're using tf.get_default_graph somewhere

lapis sequoia Dec 9, 2020, 3:48 AM

#

im not

austere swift Dec 9, 2020, 3:48 AM

#

that's deprecated as of tensorflow 2.0 so it would be tf.compat.v1.get_default_graph

lapis sequoia Dec 9, 2020, 3:48 AM

#

Traceback (most recent call last):
  File "E:/PyCharm/PYTHON projects/Pokeguesser/Pokeguesser v2/0-151/AI.py", line 109, in <module>
    model = Sequential()
  File "C:\Users\Diego\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\engine\sequential.py", line 87, in __init__
    super(Sequential, self).__init__(name=name)
  File "C:\Users\Diego\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\legacy\interfaces.py", line 91, in wrapper
    return func(*args, **kwargs)
  File "C:\Users\Diego\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\engine\network.py", line 96, in __init__
    self._init_subclassed_network(**kwargs)
  File "C:\Users\Diego\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\engine\network.py", line 294, in _init_subclassed_network
    self._base_init(name=name)
  File "C:\Users\Diego\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\engine\network.py", line 109, in _base_init
    name = prefix + '_' + str(K.get_uid(prefix))
  File "C:\Users\Diego\AppData\Local\Programs\Python\Python38\lib\site-packages\keras\backend\tensorflow_backend.py", line 74, in get_uid
    graph = tf.get_default_graph()
AttributeError: module 'tensorflow' has no attribute 'get_default_graph'

austere swift Dec 9, 2020, 3:48 AM

#

it could be that plaidml isnt updated for 2.0

lapis sequoia Dec 9, 2020, 3:48 AM

#

my tf is 2.0

#

i think (?)

#

tensorflow-estimator   2.3.0```

austere swift Dec 9, 2020, 3:49 AM

#

are you doing import keras or are you using tf.keras instead

lapis sequoia Dec 9, 2020, 3:49 AM

#

from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D```

austere swift Dec 9, 2020, 3:49 AM

#

yeah use tf.keras instead

#

so like

from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Flatten
from tensorflow.keras.layers import Conv2D, MaxPooling2D

lapis sequoia Dec 9, 2020, 3:50 AM

#

cuz i am also doing

#

import keras

austere swift Dec 9, 2020, 3:50 AM

#

yeah don't use that

#

its better to use the keras from inside tensorflow

lapis sequoia Dec 9, 2020, 3:51 AM

#

then, why did i do pip install keras?

austere swift Dec 9, 2020, 3:51 AM

#

so do from tensorflow import keras

lapis sequoia Dec 9, 2020, 3:51 AM

#

why tho?

austere swift Dec 9, 2020, 3:51 AM

#

keras actually stopped using their own functions within themselves, now it's mostly pointers just to tf.keras

lapis sequoia Dec 9, 2020, 3:51 AM

#

all the sites use keras

#

https://machinelearningmastery.com/tutorial-first-neural-network-python-keras/

Machine Learning Mastery

Jason Brownlee

Your First Deep Learning Project in Python with Keras Step-By-Step

Keras Tutorial: Keras is a powerful easy-to-use Python library for developing and evaluating deep learning models. Develop Your First Neural Network in Python With this step by step Keras Tutorial!

#

they dont use tensorflow

austere swift Dec 9, 2020, 3:52 AM

#

i mean it should theoretically work both ways

#

but tf.keras is "better" to use

#

make sure your keras is up to date as well

lapis sequoia Dec 9, 2020, 3:53 AM

#

Keras-Applications     1.0.8
Keras-Preprocessing    1.1.2```

austere swift Dec 9, 2020, 3:54 AM

#

iirc the latest version is 2.4.3

#

i'll have to double check

#

yeah it is 2.4.3

#

PS C:\Users\user> pip install --upgrade keras
Requirement already satisfied: keras in c:\users\user\appdata\local\programs\python\python38\lib\site-packages (2.4.3)

lapis sequoia Dec 9, 2020, 3:56 AM

#

aaaaaaa

#

look

austere swift Dec 9, 2020, 3:56 AM

#

and yeah thats your issue

lapis sequoia Dec 9, 2020, 3:56 AM

#


We recommend you use --use-feature=2020-resolver to test your packages with the new resolver before it becomes the default.

plaidml-keras 0.7.0 requires keras==2.2.4, but you'll have keras 2.4.3 which is incompatible.```

#

last line

austere swift Dec 9, 2020, 3:57 AM

#

ahh

lapis sequoia Dec 9, 2020, 3:57 AM

#

also, whats the error?

#

wtf

austere swift Dec 9, 2020, 3:57 AM

#

i tested using 2.2.4 and running the same model = keras.models.Sequential() line and it came up with the same error as you had

#

the tf.get_default_graph error

#

so it was the keras version causing that

lapis sequoia Dec 9, 2020, 3:58 AM

#

ERROR: After October 2020 you may experience errors when installing or updating packages. This is because pip will change the way that it resolves dependency conflicts.

#

whats this

austere swift Dec 9, 2020, 3:58 AM

#

there was a new dependency resolver implemented into pip in the latest update

#

20.3 was the version iirc

lapis sequoia Dec 9, 2020, 3:58 AM

#

so upgrading pip will fix that?

austere swift Dec 9, 2020, 3:58 AM

#

thats not an error

lapis sequoia Dec 9, 2020, 3:59 AM

#

well, it sais error

#

xDDD

austere swift Dec 9, 2020, 3:59 AM

#

thats just a warning saying itll change

lapis sequoia Dec 9, 2020, 3:59 AM

#

in capitals

austere swift Dec 9, 2020, 3:59 AM

#

it's not an actual error idk why they said that lol

lapis sequoia Dec 9, 2020, 3:59 AM

#

XD

#

okey okey, thanks for the help with keras

#

^^

austere swift Dec 9, 2020, 3:59 AM

#

Np

lapis sequoia Dec 9, 2020, 3:59 AM

#

if i knew 3060ti would be 400€ i would have saved money

#

hahahaha

austere swift Dec 9, 2020, 4:00 AM

#

Lol

lapis sequoia Dec 9, 2020, 4:00 AM

#

less complications. Cuda and going

#

😛

austere swift Dec 9, 2020, 4:00 AM

#

I'm getting a 3080 whenever they come back in stock

#

I have a 2080 super rn

#

lol

lapis sequoia Dec 9, 2020, 4:00 AM

#

i mean, i found rx580 for 100€

#

was at really good price (from 2018)

austere swift Dec 9, 2020, 4:01 AM

#

what impressed me today was that the 3080 went in stock for a whole 52 minutes

#

because they released the stock at small bursts

lapis sequoia Dec 9, 2020, 4:01 AM

#

well, what do u expect XD

austere swift Dec 9, 2020, 4:01 AM

#

but i didnt have my money ready at the time lol

lapis sequoia Dec 9, 2020, 4:01 AM

#

same numbers as 2080ti for less than half money

#

😄

austere swift Dec 9, 2020, 4:01 AM

#

more than 2080ti lol

lapis sequoia Dec 9, 2020, 4:01 AM

#

ye

austere swift Dec 9, 2020, 4:02 AM

#

3070 is the one thats comparing to 2080ti

lapis sequoia Dec 9, 2020, 4:02 AM

#

and 60ti to super

austere swift Dec 9, 2020, 4:02 AM

#

but yeah usually its gone within 10 minutes

lapis sequoia Dec 9, 2020, 4:02 AM

#

ye

austere swift Dec 9, 2020, 4:02 AM

#

yeah kinda bummed my gpu that i got 2 years ago for $750 is now beat by one thats $400

lapis sequoia Dec 9, 2020, 4:02 AM

#

xDDD

#

beat by far

austere swift Dec 9, 2020, 4:03 AM

#

at least i could probably get like 500 if i sold it since the 30 series ones are completely impossible to buy

lapis sequoia Dec 9, 2020, 4:03 AM

#

hopefully

blazing bridge Dec 9, 2020, 4:33 AM

#

Hi guys, I am building a deep learning PC and I wanted your opinion on the parts I have chosen.

i9 9900k
RTX 3070
16 GB CL16
MSI MAG 360R AIO
MSI Z390-A PRO
1TB Western Digital M.2 NVME SSD
750 W power supply
Corsair iCUE 220T RGB Airflow ATX Mid Tower Case

#

This will also be used for gaming and other productivity tasks as well

lapis sequoia Dec 9, 2020, 4:34 AM

#

ryzen is better

blazing bridge Dec 9, 2020, 4:35 AM

#

That's what they all say but intel smacks in single core performance

#

https://cpu.userbenchmark.com/Compare/Intel-Core-i9-9900K-vs-AMD-Ryzen-9-3900X/4028vs4044

UserBenchmark: AMD Ryzen 9 3900X vs Intel Core i9-9900K

Based on 470,163 user benchmarks for the AMD Ryzen 9 3900X and the Intel Core i9-9900K, we rank them both on effective speed and value for money against the best 1,276 CPUs.

lapis sequoia Dec 9, 2020, 4:36 AM

#

iirc new 5000 series are better on single threat now

blazing bridge Dec 9, 2020, 4:36 AM

#

yeah ik

#

but they are out of stock and they are a little too expensive

#

I am getting the i9 9900k for 440 CAD

lapis sequoia Dec 9, 2020, 4:38 AM

#

and 3070 isnt? XD

blazing bridge Dec 9, 2020, 4:38 AM

#

no i got it

lapis sequoia Dec 9, 2020, 4:38 AM

#

damn u lucky

blazing bridge Dec 9, 2020, 4:38 AM

#

yeah ik

lapis sequoia Dec 9, 2020, 4:38 AM

#

u should wait anyway

#

if u are not in hurry

blazing bridge Dec 9, 2020, 4:38 AM

#

im in a hurry tbh

#

i need it built by this christmas break

#

or I would have waited

lapis sequoia Dec 9, 2020, 4:39 AM

#

well, they google something like amd vs intel machine learning

#

or something

blazing bridge Dec 9, 2020, 4:41 AM

#

It’s mixed opinions

lapis sequoia Dec 9, 2020, 4:42 AM

#

graphs arent opinions. choose what u feel, im going sleep. gn

twilit brook Dec 9, 2020, 4:57 AM

#

hollow gull <@!657451103943917578> ask your questions in the chat and people will respond to...

Will do 👍

trim oar Dec 9, 2020, 6:10 AM

#

Hi guys, does "unit testing" have a special meaning in programming other than hyperparameter?

austere swift Dec 9, 2020, 7:10 AM

#

blazing bridge https://cpu.userbenchmark.com/Compare/Intel-Core-i9-9900K-vs-AMD-Ryzen-9-3900X/4...

Userbenchmark is a very unreliable source lol

austere swift Dec 9, 2020, 7:14 AM

#

blazing bridge Hi guys, I am building a deep learning PC and I wanted your opinion on the parts...

The 9900k is kinda outdated tbh, it would be better for you to go with a 10th gen cpu so you can also have further upgrade ability or to just go with one of the ryzen 5000 cpus, also for deep learning in my experience the gpu is much more important than the cpu and vram is one of my biggest bottlenecks so you might be better off getting a 2080ti for cheap on eBay or something because of the extra vram, or upgrading to a 3080 with some downgraded other parts. I’m assuming the ssd you’re talking about is the SN550, which isn’t really that good since it used HMB rather than a DRAM cache which makes it slightly slower, and also the AIO isn’t really a good value lol

#

As you might be able to tell I’m really experienced in this computer kinda stuff 🤣

#

Also btw single core doesn’t matter that much in deep learning, multi core matters much more

#

Since the main things the cpu is gonna be doing is the preprocessing, inference, and managing the data during training, all of which are highly multi threaded tasks

#

And you might want more ram if you’re gonna tackle larger datasets 😉

#

The only real performance uplift for deep learning that Intel has is AVX512, which isn’t even on the consumer chipsets only x series HEDT and the lga3647 server cpus

final scaffold Dec 9, 2020, 7:37 AM

#

data['A'] contains string like this 
{125|abs_sjowkd,
Hdujdj_hshjs,
Abs|hdus_isos_jdisi,
}

I wanna split the string on "|" (if exists) and then extract the string before the first occurrence of "_"
In a new column```

#

@trim oar sorry for the ping bro. Could you take a look at above? It's urgent.

trim oar Dec 9, 2020, 7:50 AM

#

final scaffold <@680795279758327819> sorry for the ping bro. Could you take a look at above? It...

I'm about to sleep but: https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Series.str.split.html

#

This allows you to string split and

#

expand = True would create new columns

#

I suggest playing with Regex if _ is not recognized or something

#

That should do for you

final scaffold Dec 9, 2020, 7:52 AM

#

Alright! Thanks man

azure stump Dec 9, 2020, 8:33 AM

#

https://asr373.medium.com/there-is-someone-opposite-of-sql-c15a40e55548

Medium

There is someone opposite of SQL.

Someone who can handle data without tables.

cedar yacht Dec 9, 2020, 9:08 AM

#

What model should I use for EMNIST?

#

I want to use Keras/Tensorflow to parse a single character

#

ResNet would be too overkill for this, right?

cedar yacht Dec 9, 2020, 11:34 AM

#

Wtf

pale thunder Dec 9, 2020, 11:36 AM

#

!warn @dreamy wraith do not post random gifs, especially into on topic channels

arctic wedgeBOT Dec 9, 2020, 11:36 AM

#

:incoming_envelope: :ok_hand: applied warning to @dreamy wraith.

torpid cave Dec 9, 2020, 11:39 AM

#

I think I will start moving towards NoSQL tables

#

I just did a quick course on it and they seem much easier to manage for my kind of work

jolly zealot Dec 9, 2020, 11:41 AM

#

Hi!

It's Akshay.
I am looking for full-time 'Python' or 'Artificial Intelligence' pair programmers. If you would like to pair-program, please message your availability and schedule.

Thank you.

Best,
Akshay.

austere swift Dec 9, 2020, 2:31 PM

#

!rule 6

arctic wedgeBOT Dec 9, 2020, 2:31 PM

#

Rules

6. No spamming or unapproved advertising, including requests for paid work. Open-source projects can be shared with others in #python-general and code reviews can be asked for in a help channel.

lapis sequoia Dec 9, 2020, 4:30 PM

#

If anyone is familiar with the SciPy solve_ivp function, I could use some help on how to implement it for a system of ODEs: https://scicomp.stackexchange.com/questions/36465/solve-coupled-odes-in-time-and-space-using-python

Computational Science Stack Exchange

Solve coupled ODEs in time and space using Python

I have a system that is represented by the equations shown below. This is part of a larger system of equations but for the purpose of this question I have simplified things.
$$
\frac{\partial c}{\p...

merry wadi Dec 9, 2020, 5:19 PM

#

Is there a way to use a decision tree with a target variable that has multiple values?
Rather than predicting (0 or 1) it would predict (0,1,2,3)

upbeat bone Dec 9, 2020, 5:39 PM

#

i have 4 variables, "age" , "sex" , "year" and "expected" & am trying to plot a graph using only the data from "male" which is in the "sex" column in my .csv file .. I have tried lifedf.plot.bar('year','expected') , but this is just plotting it all

sand sluice Dec 9, 2020, 5:48 PM

#

lifedf[lifedf['sex'] == 'male'].plot.bar('year','expected') ?

lapis sequoia Dec 9, 2020, 5:51 PM

#

I posted my question in the #help-kiwi channel.

upbeat bone Dec 9, 2020, 6:12 PM

#

doesnt seem to work 💔

trim oar Dec 9, 2020, 6:27 PM

#

upbeat bone i have 4 variables, "age" , "sex" , "year" and "expected" & am trying to plot a ...

hue='sex' Edit: read it wrong. Don't have device to play right now tho

lapis sequoia Dec 9, 2020, 6:34 PM

#

I'm trying to convert my script from tensorflow v1 to tensorflow v2.
It is a frozen_inference_model, and I managed to convert it into a saved model so far.
Now when it comes to inference, in Tensorflow V1 I had to pass the value of
{self.graph.get_tensor_by_name('image_tensor:0'):data}

I'm unsure how to do this in Tensorflow V2 with a savedmodel, on how to get the tensor by name.
Help is appreciated

fair kernel Dec 9, 2020, 7:32 PM

#

hello guys im new in data science and want to evolve my skils fast and good

#

im planning to start a twitch channel for data science learning

#

starting wiht 30days of code python for data science

#

you think tahts a good idea?

#

or is this what people wnat to watch ? 😄 ❤️ thanks in advance guys

#

and girls 🙂

fading wigeon Dec 9, 2020, 7:55 PM

#

Anyone use databricks before? How hard is it?

trim oar Dec 9, 2020, 8:01 PM

#

fair kernel hello guys im new in data science and want to evolve my skils fast and good

what's your goal here... learn or get audience

next talon Dec 9, 2020, 8:35 PM

#

How can I reduce my validation loss ?

fair kernel Dec 9, 2020, 9:20 PM

#

@muted thorn yu learn! best way of learning is to teach ohters while you learning

#

@trim oar

#

@trim oar think about the rubberduck method in software development

#

what you want to achive?

fair kernel Dec 9, 2020, 9:23 PM

#

trim oar what's your goal here... learn or get audience

iam just feeling that new learners have a hard time getting into things, i am enrolled in udacity nanodegree for data anylst and most guys there need to research alot on google. cuz they beginners in programming at all

azure hearth Dec 9, 2020, 9:31 PM

#

@ gzuma , im not a beginner and still search everything on google 😛

#

sry 😉

fair kernel Dec 9, 2020, 9:31 PM

#

azure hearth @ gzuma , im not a beginner and still search everything on google 😛

yes yes me too hahha, but you know how stuff can work, beginners dont konw how stuff can work and they have to search hours and hours for it

azure hearth Dec 9, 2020, 9:32 PM

#

actually not really^^

fair kernel Dec 9, 2020, 9:32 PM

#

i just watched yt videos and blgo post on medium for useufll trick

azure hearth Dec 9, 2020, 9:32 PM

#

but python is really easy

fair kernel Dec 9, 2020, 9:32 PM

#

for someone who never has programmed or know how digital stuff is working its hard to learn python!^^

azure hearth Dec 9, 2020, 9:33 PM

#

try learn c++ 🙂

#

or assembler

fair kernel Dec 9, 2020, 9:33 PM

#

no way 😄 haha

#

why i should learn c++`? im learnign data science just to get the power of data insights

#

and building nice products

#

yea youre right c++ would e helpful 😄

#

be*

azure hearth Dec 9, 2020, 9:34 PM

#

you wont need to learn python if you can do c++ and java

#

i never learned python

fair kernel Dec 9, 2020, 9:35 PM

#

i know

#

but im ont only learning programming

#

i dont want to be an programmer

azure hearth Dec 9, 2020, 9:35 PM

#

what you do exactly?

fair kernel Dec 9, 2020, 9:36 PM

#

learning python ds, ux, building companys products, marketing

#

everythign that is necessary to bring ideas to life!

azure hearth Dec 9, 2020, 9:36 PM

#

cool

fair kernel Dec 9, 2020, 9:37 PM

#

At some point is rly helpful to make a MVP of an idea or product

azure hearth Dec 9, 2020, 9:37 PM

#

im just doing websites to earn money

fair kernel Dec 9, 2020, 9:37 PM

#

ye this is great ❤️

#

would love to only do websites 🙂

azure hearth Dec 9, 2020, 9:37 PM

#

just do it 😛

fair kernel Dec 9, 2020, 9:37 PM

#

TiME

#

day only have 24 hours

#

im more passionate about doing my own stuff then doing clients stuff

azure hearth Dec 9, 2020, 9:38 PM

#

money ! 😄

fair kernel Dec 9, 2020, 9:39 PM

#

you earning millions?

azure hearth Dec 9, 2020, 9:39 PM

#

no, but i also dont need millions

fair kernel Dec 9, 2020, 9:40 PM

#

send me portfolio as PM

azure hearth Dec 9, 2020, 9:40 PM

#

everything should have website for frontend stuff , i think

fair kernel Dec 9, 2020, 9:40 PM

#

website is basic marketing stuff

azure hearth Dec 9, 2020, 9:40 PM

#

portfolio from me ?, im not searching for jobs

fair kernel Dec 9, 2020, 9:41 PM

#

im just an design lover ❤️ and love creative work

#

website portfolio just to watch what i will never can achieve : D

azure hearth Dec 9, 2020, 9:42 PM

#

stuff like that

#

thats a fair website

trim oar Dec 9, 2020, 9:51 PM

#

@fair kernel so it’d be streaming about you going through the udacity course.

fair kernel Dec 9, 2020, 9:51 PM

#

@trim oar i thaught about somethign like that

#

i have a data science project ongoing

#

gather spotify data

#

analysie it

trim oar Dec 9, 2020, 9:52 PM

#

I’d say go for it

fair kernel Dec 9, 2020, 9:52 PM

#

buildign pipeline

#

build insigt plattform

trim oar Dec 9, 2020, 9:52 PM

#

I just don’t know if Twitch has the right audience

fair kernel Dec 9, 2020, 9:52 PM

#

trim oar I just don’t know if Twitch has the right audience

but which plattform have it, i created a tiktok accoutn already

trim oar Dec 9, 2020, 9:53 PM

#

I don’t watch stream, so honestly no clue. Considering more people search YouTube videos, I’d imagine if not now, at least those VODs would be useful later

fair kernel Dec 9, 2020, 9:55 PM

#

trim oar I don’t watch stream, so honestly no clue. Considering more people search YouTub...

was thinking about yt too, maybe for udacity nanodegree tips and streams for interaction with the community ❤️

#

i just have no time for produce yt videos 😦

trim oar Dec 9, 2020, 9:58 PM

#

Honestly there are plenty of unedited stuff on there that gained popularity

#

Your goal was to learn data science anyways

#

YT also has streaming function btw

#

Or you can stream both at the same time

#

If your computer can handle it

#

And see which one performs

fair kernel Dec 9, 2020, 10:09 PM

#

trim oar Honestly there are plenty of unedited stuff on there that gained popularity

youre right 🙂 i have bad habit to always want to make high quaitly work

#

😦

fair kernel Dec 9, 2020, 10:12 PM

#

trim oar Honestly there are plenty of unedited stuff on there that gained popularity

was thinkg about yt series, short clisp 1-2 minutes about easy steps for beginners like 2 minute tutorial how to plot countplots

trim oar Dec 9, 2020, 10:13 PM

#

As long as you won’t be too hung up by the numbers

fair kernel Dec 9, 2020, 10:14 PM

#

prjoct based learnign is a thing , so every video will connect to 1 big proeject useing the projects data to make the little videos

trim oar Dec 9, 2020, 10:14 PM

#

Because I’ve seen great talents who later on got too caught up by the underperforming numbers.

fair kernel Dec 9, 2020, 10:14 PM

#

you mean views

azure hearth Dec 9, 2020, 10:14 PM

#

noone want to watch programming lessons :/

trim oar Dec 9, 2020, 10:14 PM

#

Yup

#

Because for example, YouTube recommends videos based on a few metrics, and 1-2 minute would be ranked lower than 10 minute videos even if it’s the “freshness” is the same

fair kernel Dec 9, 2020, 10:15 PM

#

https://www.youtube.com/watch?v=XezfbWlEVwE

YouTube

Krish Naik

Tutorial 10- Seaborn- Countplot(), Violinplot(), Boxplot()- Part2

Hello All,
Welcome to the Python Crash Course. In this video we will understand about Seaborn

github url : https://github.com/krishnaik06/Machine-Learning-in-90-days

Support me in Patreon: https://www.patreon.com/join/2340909?

Connect with me here:
Twitter: https://twitter.com/Krishnaik06
Facebook: https://www.facebook.com/krishnaik06
instagr...

▶ Play video

trim oar Dec 9, 2020, 10:16 PM

#

If it got viral for random luck however it won’t stop recommending

fair kernel Dec 9, 2020, 10:16 PM

#

22k views 😄

trim oar Dec 9, 2020, 10:16 PM

#

I know it’s a thing.

azure hearth Dec 9, 2020, 10:16 PM

#

22k = 22€ ?

fair kernel Dec 9, 2020, 10:17 PM

#

22.000 views

trim oar Dec 9, 2020, 10:17 PM

#

But it was built over a year with the right keywords, and if you got too caught up with numbers, especially fresh channels

fair kernel Dec 9, 2020, 10:17 PM

#

this is an data anylsi sproject guys, find out what average duration a video have in yt, realted to programming lessons videos 😄

trim oar Dec 9, 2020, 10:17 PM

#

Then one might be very disappointed

fair kernel Dec 9, 2020, 10:17 PM

#

its a figth

#

to gain views

#

keywods, audio quality, title and thubmnails

#

shoudl do the work

trim oar Dec 9, 2020, 10:18 PM

#

And this guy has 225k subscribers with years of building the channel

#

No

#

YouTube algorithm is a thing

#

Same with Twitch and other platforms

azure hearth Dec 9, 2020, 10:18 PM

#

maybe you can get informations with the youtube api

trim oar Dec 9, 2020, 10:18 PM

#

The amount of work going into those design is insane.

fair kernel Dec 9, 2020, 10:19 PM

#

right 🙂 youtube api then doing research

#

have 3 yeras of design school^^

#

ahhh

#

ahah

#

you meant the youtube algo

trim oar Dec 9, 2020, 10:19 PM

#

Yes

fair kernel Dec 9, 2020, 10:19 PM

#

haha

#

😄

trim oar Dec 9, 2020, 10:19 PM

#

It’s recommendation also

#

Algo

fair kernel Dec 9, 2020, 10:19 PM

#

yeye i know

trim oar Dec 9, 2020, 10:19 PM

#

Its

fair kernel Dec 9, 2020, 10:19 PM

#

it is

#

trim oar Dec 9, 2020, 10:20 PM

#

Yeah no what I simply meant is as long as you don’t sway your main goal of learning data science because of the numbers then I think it’s great

fair kernel Dec 9, 2020, 10:20 PM

#

there shoudl be an webservice which is connected to youtube api to find out which things you have to do to get attraction to viewers, by category and keywords

trim oar Dec 9, 2020, 10:20 PM

#

But if worrying about the numbers decrease your quality of learning then it’s a bad idea

#

No it’s more complicated than that

fair kernel Dec 9, 2020, 10:20 PM

#

nah nohting can bring my motivation down 🙂

trim oar Dec 9, 2020, 10:21 PM

#

You see it’s not key words based. That’s like more than five years ago

#

And it’s updating every year

#

Currently primarily based on average watch time, so even if people finish 100% of 2 minute video and 50% of 10 minute video, theirs still twice more than likely to get recommended

#

But there are other things, like kids related is more punished now because of the privacy issues

fair kernel Dec 9, 2020, 10:23 PM

#

i mean a service for end users. they type in somem keywords and get recomendation data or analysis, from data of youtube api,
then they see what things succesufl videos., related to their key word search, have.
e.g like thumbnails, taging, vidoe lenght, countrys etc .....

trim oar Dec 9, 2020, 10:24 PM

#

TubeBuddy more or less does it

#

But it’s also personalized

#

Highly personalized

fair kernel Dec 9, 2020, 10:26 PM

#

trim oar TubeBuddy more or less does it

will check it out, thx ❤️

trim oar Dec 9, 2020, 10:27 PM

#

🥰

fair kernel Dec 9, 2020, 10:27 PM

#

other qestion. you worked with selenium?

trim oar Dec 9, 2020, 10:28 PM

#

No unfortunately

fair kernel Dec 9, 2020, 10:28 PM

#

i made scraper for creating pandas dataframes but selenium run to long

#

😦 ok ok

trim oar Dec 9, 2020, 10:29 PM

#

I’ve only worked with BeautifulSoup for webscraping before

#

But not the same thing

fair kernel Dec 9, 2020, 10:32 PM

#

yes, did this too but beatifulsoup didnt worked for the spotify web page ,
only selenium because it works with the javascript website stuff

#

https://www.kaggle.com/c/kaggle-survey-2020

2020 Kaggle ML & DS Survey

The most comprehensive dataset available on the state of ML and data science

#

everony have already a team on it ? 30,000 dollars seems attractive

blazing bridge Dec 9, 2020, 10:50 PM

#

I don't have a team but I wouldn't mind trying if someone is interested.

fair kernel Dec 9, 2020, 11:02 PM

#

great

#

you already did a copmetition?

#

@blazing bridge ?

blazing bridge Dec 9, 2020, 11:13 PM

#

no i havent done a competition

#

yet

odd jackal Dec 10, 2020, 4:45 AM

#

hii
does anyone know how do i get kaggle_secrets to work in jupyter notebook?

ripe lion Dec 10, 2020, 8:00 AM

#

Hello, So I'm working with this data frame:
https://cdn.discordapp.com/attachments/303906576991780866/785897567317131295/Capture2.PNG

#

I'm trying to compare the votes for republicans and democrats to then plot them on a map but I'm stuck.
Here's the code I came up with: https://cdn.discordapp.com/attachments/430969524011925507/786503312383148042/Capture.PNG

#

but here is the result: https://cdn.discordapp.com/attachments/303906576991780866/785897469098721350/Capture3.PNG

#

Anybody knows how I can get rid of the "NaN" thing on 'democrat_votes' and 'republican_votes' and merge them so that each state is in one row?

sweet plaza Dec 10, 2020, 8:32 AM

#

@ripe lion Tru using this on Pandas https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.dropna.html

#

the dropna function

ripe lion Dec 10, 2020, 8:57 AM

#

oh thank you @sweet plaza , will try that!

velvet thorn Dec 10, 2020, 9:46 AM

#

ripe lion oh thank you <@!406014882673328128> , will try that!

use combine_first

#

df['republican_votes'].combine_first(df['democrat_votes'])

velvet thorn Dec 10, 2020, 9:46 AM

#

sweet plaza the dropna function

dropna is going to remove everything

ripe lion Dec 10, 2020, 10:39 AM

#

velvet thorn `df['republican_votes'].combine_first(df['democrat_votes'])`

Oh, Thanks a lot!

velvet thorn Dec 10, 2020, 10:48 AM

#

<@&267629731250176001> spam

eager heath Dec 10, 2020, 10:49 AM

#

!tempban @wanton bison 14d It seems like you're only here to spam a link to your course. Re-read our rules if you want to be part of this server.

arctic wedgeBOT Dec 10, 2020, 10:49 AM

#

:incoming_envelope: :ok_hand: applied ban to @wanton bison until 2020-12-24 10:49 (13 days and 23 hours).

torpid cave Dec 10, 2020, 10:58 AM

#

sup you all

lilac kindle Dec 10, 2020, 10:58 AM

#

what is a smart way to check if values in a list are increasing or decreasing?

torpid cave Dec 10, 2020, 11:00 AM

#

@lilac kindle visually?

#

Maybe doing an if condition and prompting a message

lilac kindle Dec 10, 2020, 11:00 AM

#


[1 3 7 8]  #  all 4 values increasing
[11 9 7 2]  #  all 4 values decreasing
[1 11 7 8]  # last 2 values increasing

torpid cave Dec 10, 2020, 11:00 AM

#

Oh so you mean a static list

#

Is this inside a loop?

lilac kindle Dec 10, 2020, 11:01 AM

#

wait lemme fix the format it's irritating

#

I thought that was the command

torpid cave Dec 10, 2020, 11:02 AM

#

You could use numpy to do the comparison

#

e.g.

import numpy as np
np.array([list0]) > np.array(list1])

#

And then count true/false

#

Use indexes if you want to know exactly what values increase/decrease

#

output = np.array([1, 2, 3, 4]) > np.array([2, 3, 4, 5])
print(sum(output), "values are increasing")

#

Something among those lines

#

It tells you the number

#

Then you could custom use the print function right

#

add some conditionals to get different messages

lilac kindle Dec 10, 2020, 11:14 AM

#

I don't understand

#

I don't rly want to compare two lists with eachother, just find a pattern in a single list, like if the last 10 values were constantly increasing: do something

torpid cave Dec 10, 2020, 11:15 AM

#

You need to store 2 lists...

#

The first list would be the old version, the second the new version

#

My guess is that you are doing this inside a loop/function

lilac kindle Dec 10, 2020, 11:18 AM

#

yea

#

I'm constantly adding values so the dimensions are off

#

new list will have a new extra value

torpid cave Dec 10, 2020, 11:20 AM

#

Try to do this... at the beginning of the loop save your list in a value, then when you transform it, at the end of the loop, run your conditional

#

And print what is happening

#

If the data is not too large

lilac kindle Dec 10, 2020, 11:20 AM

#

ye true

torpid cave Dec 10, 2020, 11:20 AM

#

Maybe just create a list of lists

#

For every iteration

#

And try to see it graphically

lilac kindle Dec 10, 2020, 11:22 AM

#

thanks for the help! got plenty of stuff to try now

torpid cave Dec 10, 2020, 11:22 AM

#

No worries!

hasty grail Dec 10, 2020, 11:24 AM

#

!e

def is_nondecreasing(lst):
    if len(lst) <= 1:
        return True

    prev = lst[0]
    for e in lst:
        diff = e - prev
        if diff < 0:
            return False
        prev = e

    return True

print(is_nondecreasing([1, 5, 5, 9, 2]))
print(is_nondecreasing([1, 5, 5, 9, 10]))

arctic wedgeBOT Dec 10, 2020, 11:25 AM

#

@hasty grail :white_check_mark: Your eval job has completed with return code 0.

001 | False
002 | True

hasty grail Dec 10, 2020, 11:25 AM

#

This will work for a simple list, if you have a nested list you'll have to do it in a loop

torpid cave Dec 10, 2020, 11:25 AM

#

Easier/cleaner solution

hasty grail Dec 10, 2020, 11:26 AM

#

I'm assuming that you're not using numpy

#

!e

import numpy as np

def is_nondecreasing(lst):
    return (np.diff(lst) >= 0).all()

print(is_nondecreasing([1, 5, 5, 9, 2]))
print(is_nondecreasing([1, 5, 5, 9, 10]))

arctic wedgeBOT Dec 10, 2020, 11:28 AM

#

@hasty grail :white_check_mark: Your eval job has completed with return code 0.

001 | False
002 | True

hasty grail Dec 10, 2020, 11:28 AM

#

Otherwise you can just do this ^

torpid cave Dec 10, 2020, 11:30 AM

#

Any way I could make this a one-liner

#

data['AttractiveA'] = 8 - data['AttractiveA']
data['AttractiveB'] = 8 - data['AttractiveB']
data['AttractiveC'] = 8 - data['AttractiveC']

hasty grail Dec 10, 2020, 11:35 AM

#

Using a list comp, yes, but it's bad practice to have side effects in list comps. You should use a for loop instead

#

for k in ('AttractiveA', 'AttractiveB', 'AttractiveC'):
    data[k] = 8 - data[k]

torpid cave Dec 10, 2020, 11:37 AM

#

That was quite simple

#

Thanks

hasty grail Dec 10, 2020, 11:37 AM

#

np

torpid cave Dec 10, 2020, 11:37 AM

#

I am transforming a dataframe btw

#

I will actually just use that to transform all the columns

hasty grail Dec 10, 2020, 11:37 AM

#

hmm you might be able to get away using multiindexing

torpid cave Dec 10, 2020, 11:39 AM

#

data['AttractiveA'] = 8 - data['AttractiveA']
data['AttractiveB'] = 8 - data['AttractiveB']
data['AttractiveC'] = 8 - data['AttractiveC']
    
#Inverse Innovation
data['InnovativeA'] = 8 - data['InnovativeA']
data['InnovativeB'] = 8 - data['InnovativeB']
data['InnovativeC'] = 8 - data['InnovativeC']    
    
#Inverse Easy
data['EasyuseA'] = 8 - data['EasyuseA']
data['EasyuseB'] = 8 - data['EasyuseB']
data['EasyuseC'] = 8 - data['EasyuseC']    

# plus 8 additional variables

#

I was thinking using something like... starts_with (R based)

#

I think it is cleaner to just create a list of the columns and loop over it

hasty grail Dec 10, 2020, 11:41 AM

#

Not that experienced in pandas, unfortunately

torpid cave Dec 10, 2020, 11:41 AM

#

No worries, me neither

#

I try to avoid them as much as possible

hasty grail Dec 10, 2020, 11:42 AM

#

I play around with numpy a lot but not really pandas

torpid cave Dec 10, 2020, 11:43 AM

#

Well this are survey results so I guess it makes sense using pandas

#

*these

hasty grail Dec 10, 2020, 11:43 AM

#

Yeah

torpid cave Dec 10, 2020, 11:43 AM

#

But I do all the matrix calculations with np though

#

@hasty grail want to try a slightly more challenging one? I optimized it already but maybe you could think about a different way of doing it

hasty grail Dec 10, 2020, 11:52 AM

#

Sure

torpid cave Dec 10, 2020, 11:53 AM

#

So I am doing this:

AppealA = np.nanmean(np.array([data['AttractiveA'],data['WowA'],data['LoveA']]),axis=0)
AppealB = np.nanmean(np.array([data['AttractiveB'],data['WowB'],data['LoveB']]),axis=0)
AppealC = np.nanmean(np.array([data['AttractiveC'],data['WowC'],data['LoveC']]),axis=0)

#

But like Appeal, I have about 20 other variables

#

My quick solution was

#

for factor in [A, B, C]:
  var = 'Appeal' + factor
  x = 'Attractive' + factor
  y = 'Wow' + factor
  z = 'Love' + factor

  data[var] = np.nanmean(np.array([data[x],data[y],data[z]]), axis=0)

  var2 = 'Var2' + factor
  #x = .....

#.... up to Var 20

#

opps

#

And note that some variables have 3 inputs, some have up to 8 inputs

#

Don't want to make it too complicated though as I have to pass this code to someone else to maintain

hasty grail Dec 10, 2020, 11:58 AM

#

appeals = {}
bases = ('Attractive', 'Wow', 'Love')
for subcol in ('A', 'B', 'C'):
    cols = [base + subcol for base in bases]
    appeals['Appeal' + subcol] = np.nanmean(data.loc[:, cols].to_numpy())

#

Not entirely sure this would work but worth a try xD

torpid cave Dec 10, 2020, 12:00 PM

#

hahahaha

#

Well I just realized both solutions are longer than 3 lines

#

I might turn yours into a function

hasty grail Dec 10, 2020, 12:01 PM

#

We're not code golfing so usually it would be longer than 3 lines, lol

torpid cave Dec 10, 2020, 12:02 PM

#

Oh I mena

#

mean

#

The basic solution was 3 lines

#

ugly but 3 lines

#

Is fun doing this though, I should focus a bit more on finishing my work though

hasty grail Dec 10, 2020, 12:02 PM

#

You should try to follow the DRY principle as much as possible

#

Otherwise it will become more difficult to maintain the code in the future

torpid cave Dec 10, 2020, 12:03 PM

#

true

#

@hasty grail

def ComputeScore(score_name, components):
        result = {}
        for prod in ('A', 'B', 'C'):
            cols = [comp + prod for comp in components]
            result[score_name + prod] = np.nanmean(data.loc[:, cols].to_numpy(), axis=0)
        return result

#

x = ComputeScore('Appeal',['Attractive','Wow','Love'])

#

nvm

#

Working at 90%

#

I get results rather than matriz

#

a matrix

hasty grail Dec 10, 2020, 12:11 PM

#

you can convert that easily

velvet thorn Dec 10, 2020, 12:11 PM

#

@torpid cave store the column names in some data structure

#

then you can access the subset of the DF corresponding to them directly

#

and you don't need to use np.nanmean

#

normal pandas .mean will work just fine

#

looping is inefficient

hasty grail Dec 10, 2020, 12:13 PM

#

^ listen to this pandas expert xD

torpid cave Dec 10, 2020, 12:14 PM

#

pandas takes care of NAs?

velvet thorn Dec 10, 2020, 12:14 PM

#

!e

import pandas as pd

print(pd.Series([1, float('nan'), 2]).mean())

arctic wedgeBOT Dec 10, 2020, 12:14 PM

#

@velvet thorn :white_check_mark: Your eval job has completed with return code 0.

1.5

torpid cave Dec 10, 2020, 12:15 PM

#

Ok it does

#

so I will just use this


for product in ('A','B','C'):
        data['Appeal' + product] = data[['Attractive' + product, 'Wow' + product, 'Love' + product]].mean(axis=1)
# State additional variables after this

#

Thanks all

velvet thorn Dec 10, 2020, 12:19 PM

#

🥴

#

well I mean that's not so bad

torpid cave Dec 10, 2020, 12:20 PM

#

I actually like it

#

it is 1 line per variable

velvet thorn Dec 10, 2020, 12:20 PM

#

yeah is fine I guess

#

I mean

#

you could optimise it but maybe not owrth

torpid cave Dec 10, 2020, 12:21 PM

#

Yep nww

#

I am just moving data from one place to another, every friday

#

For datasets that will never be considered big

#

You should have seen the initial R code I made for this 1-year ago

vagrant tulip Dec 10, 2020, 12:39 PM

#

is there an image quality metric with both rotational invariance and greyscale invariance?

fallow rune Dec 10, 2020, 1:39 PM

#

I was working with the Black Friday Dataset and got 64% accuracy on a Random Forest Regression model. What can I do to improve it?

shut apex Dec 10, 2020, 2:02 PM

#

Hi, does anyone know why there's a syntax error in this?

heat_data = [[row['Latitude'],row['Longitude']] for index, row in heat_df.iterrows()]

lapis sequoia Dec 10, 2020, 2:56 PM

#

is there a help channel for r programming

old thorn Dec 10, 2020, 2:58 PM

#

@lapis sequoia don't believe so but there are many professional data scientists in the server so if you DM one of them, they could help you

ember silo Dec 10, 2020, 3:16 PM

#

Morning all!

#

Just getting into Tensorflow and was wondering if anyone had used the Raspberry Pi hat for it

tough fiber Dec 10, 2020, 4:24 PM

#

Hi, im wondering if XOR gate can be solved with single perceptron?

#

if so, how?

livid quartz Dec 10, 2020, 4:37 PM

#

Am I using TomekLinks right ? My recall is still very low

📎 unknown.png

azure stump Dec 10, 2020, 5:58 PM

#

https://medium.com/analytics-vidhya/5-must-have-soft-skills-for-data-scientists-1952d8e5d7a

Medium

5 Must have Soft Skills for Data Scientists.

Being great at your job is more than the code.

lilac raven Dec 10, 2020, 7:03 PM

#

how would I go about closing a figure after receiving no new incoming data? I tried for val in read_fname: if val is None: plt.close() inside and outside my animation loop but I don't think it's actively looking for new data even after the seperate file that outputs data is closed. I'd guess the code will analyze if it's given new data, but how would I go about adding a line to check if stream is even open

fair kernel Dec 10, 2020, 7:34 PM

#

fallow rune I was working with the Black Friday Dataset and got 64% accuracy on a Random For...

use xgboost?

fair kernel Dec 10, 2020, 9:31 PM

#

streamlit?

outer fulcrum Dec 10, 2020, 9:34 PM

#

Hey guys ! I have data from solar installations which are stored on an FTP server (csv files). I need a dashboard to visualize metrics and a way to define alerts. What kind of tools should I use ?

fair kernel Dec 10, 2020, 9:37 PM

#

outer fulcrum Hey guys ! I have data from solar installations which are stored on an FTP serve...

streamlit?

fair kernel Dec 10, 2020, 10:17 PM

#

im working on one

#

look at streamlit homepage there are examples

#

https://www.streamlit.io/gallery

Gallery | Streamlit — The fastest way to create data apps

Browse through some of the amazing apps that the community has created with Streamlit. If you want to have your app featured, just tweet it to us @streamlit!

serene scaffold Dec 11, 2020, 2:40 AM

#

Is anyone else here interested in NLP? I'm contemplating a program that, given a series of messages in a Discord-like environment, predicts which messages, if any, are being responded to by other messages.

#

So for example, gzuma's comment "im working on one" would be predicted as a response to ragepope's comment "you have any example projects with this?"

#

My colleague suggested I look into textual entailment but I doubt there's an off-the-shelf solution without making my own training data.

sour marsh Dec 11, 2020, 3:02 AM

#

Hey guys - anyone have experience with k-means clustering

#

having some difficulty using kmeans - keep getting errors

torpid cave Dec 11, 2020, 3:08 AM

#

Hi guys

#

Hi do I increase a variable alphabetically within a loop... e.g.

for i in range(1,10):
  print(letter) #This should return 'A'
  #function to increase the letter, so the next iteration it prints 'B'

#

Somethingl ike that

#

DId not know about ascii_lowercase

#

Thank you

#

so much

#

I was googling this stuff but I didn't know how to formulate the question haha

#

actually that is simple

#

and easier

tight torrent Dec 11, 2020, 3:20 AM

#

I need some help with the Tweepy module

torpid cave Dec 11, 2020, 3:21 AM

#

I am gonna vectorize it now

tight torrent Dec 11, 2020, 3:21 AM

#

sometimes when tweepy is getting The tweet it Errors and i think its because of the characters. how do i prevent it..

📎 unknown.png

torpid cave Dec 11, 2020, 3:27 AM

#

No worries haha

lapis sequoia Dec 11, 2020, 4:49 AM

#

lapis sequoia is there a help channel for r programming

no but u could ask in offtopic

tight torrent Dec 11, 2020, 5:04 AM

#

📎 unknown.png

astral pollen Dec 11, 2020, 5:45 AM

#

@tight torrent If it is a retweet, the api returns it as truncated with ... if over a certain length. Is that what happens in your case?

tight torrent Dec 11, 2020, 5:49 AM

#

astral pollen <@429535933252239360> If it is a retweet, the api returns it as truncated with ....

its not the retweet

#

its the text

#

i saw stack posts and everyone said use the full_text arg instead of status.text

#

so i did the same

#

but still no work

#

no errrs

astral pollen Dec 11, 2020, 5:51 AM

#

hmm

#

are you just printing to stdout or how are you dealing with the tweet object?

heady hatch Dec 11, 2020, 6:05 AM

#

Hello people. Question about tf.

I'm bumping into this error but I'm confused why.

InvalidArgumentError:  Matrix size-incompatible: In[0]: [100,751], In[1]: [750,32]
     [[node functional_17/dense_18/Relu (defined at <ipython-input-91-9f049ca8f25a>:1) ]] [Op:__inference_train_function_5904]

The layers in question are

______________________________________________________________________________________________
concatenate_13 (Concatenate)    (None, 750)          0           normalization_7[0][0]            
                                                                 normalization_8[0][0]            
                                                                 category_encoding_11[0][0]       
                                                                 category_encoding_12[0][0]       
______________________________________________________________________________________________
dense_18 (Dense)                (None, 32)           24032       concatenate_13[0][0]

I'm not sure what to make of this. Because I'm under the assumption that the shape would be (None, 750), but then it's outputting (None, 751) at concatenate_13?

tight torrent Dec 11, 2020, 6:09 AM

#

oh

#

lemme se

tight torrent Dec 11, 2020, 6:11 AM

#

astral pollen are you just printing to stdout or how are you dealing with the tweet object?

im printing it to a stdout

#

yes

#

i have created a text file

#

the status will print it things

#

into the file

#

umm why u deleted the code? @astral pollen

astral pollen Dec 11, 2020, 6:13 AM

#

I am on iphone so indentation got weird. Here it is:

#

if "extended_tweet" in status_json:   
    print(status_json['extended_tweet']['full_text'])
elif 'retweeted_status' in status_json:
    if 'extended_tweet' in status_json['retweeted_status']:   
    print(status_json['retweeted_status']['extended_tweet']['full_text']) 
    else: 
        print(status_json['text']) 
else: print(status_json['text'])

#

the if x in status_json bits must be changed according to how you deal with the tweet object in your own code

tight torrent Dec 11, 2020, 6:26 AM

#

ok

#

@astral pollen

#

tweet_json not defined

#

what is it supposed to be?

astral pollen Dec 11, 2020, 6:34 AM

#

can you show your code?

tight torrent Dec 11, 2020, 6:35 AM

#

alr

#

class Tweet_analyzer():
    def tweets_to_dataframe(self, tweets):

        df = pandas.DataFrame(data=[tweet.full_text for tweet in tweets], columns=['Tweets'])

        #df['Date'] = numpy.array([tweet.created_at for tweet in tweets])
        df['Likes'] = numpy.array([tweet.favorite_count for tweet in tweets])
        df['Retweets'] = numpy.array([tweet.retweet_count for tweet in tweets])
        df['Tweet ID'] = numpy.array([tweet.id for tweet in tweets]) 

        return df



if __name__ == '__main__':
    command = input("Enter username: ")
    try:
        client = Twitter_client()
        analyzer = Tweet_analyzer()

        api = client.get_twitter_client_api()

        tweets = api.user_timeline(screen_name=command,count=20, tweet_mode = 'extended')

        df = analyzer.tweets_to_dataframe(tweets)
        #print(dir(tweets[0]))

        sys.stdout = open("Output.txt", "w",encoding='utf-8')

        print(df.head(10))

        sys.stdout.close()
    except Exception as e:
            print("User not found")
            print(e)

#

this is the main part

#

the others is just the auth and other boring stuff

#

@astral pollen

azure stump Dec 11, 2020, 7:25 AM

#

https://asr373.medium.com/without-performing-this-technique-you-cant-train-your-ml-model-9c6c3a78814f

Medium

Without performing this technique you can’t train your ML Model.

Everyone used to neglect this.

fading monolith Dec 11, 2020, 10:09 AM

#

Can anyone please suggest a list of projects to do, that I can learn ML through? I’ve tried doing courses and they are boring, and I think doing projects and learning through them will be more fun...

#

Thank in advance!

lapis sequoia Dec 11, 2020, 11:05 AM

#

In my pandas dataframe, a series say 'countries' have list of countries separated with dashes (-), how can I replace that with commas (,)?

bitter harbor Dec 11, 2020, 12:26 PM

#

fading monolith Can anyone please suggest a list of projects to do, that I can learn ML through?...

honestly i'd suggest learning the absolute fundamentals and building a model yourself without using a library (tf, keras, etc), it'll be a lot more work than following a course but imo you get more out of it + it makes the concepts easier to understand

worldly palm Dec 11, 2020, 12:47 PM

#

Anyone mind helping me out in matplotlib labeling with scatter plots? (I'm also in #help-burrito...)

fallow prism Dec 11, 2020, 2:09 PM

#

can someone explain me why this happened? yesterday i run my gensim TfIdf model and obtain vectors shwed in figure 1 and today i get figure 2, both codes are identical.

#

figure 1

📎 good_dist.png

#

ran*

earnest forge Dec 11, 2020, 2:26 PM

#

should I normalize/scale data for classification models?

#

especially, latitue and longtitute values

#

do they really need to be scaled?

fallow prism Dec 11, 2020, 2:43 PM

#

yes you have

earnest forge Dec 11, 2020, 2:45 PM

#

I suppose, it is better use StandardScaler, since there are negative values in coordinates data aswell

lapis sequoia Dec 11, 2020, 2:48 PM

#

I'm having some issues with VSCode and Miniconda on a Mac. If anyone has suggestions, please let me know in #help-avocado

fallow prism Dec 11, 2020, 3:14 PM

#

fig2

📎 good_dist_today.png

molten hamlet Dec 11, 2020, 3:16 PM

#

i know there is some geopandas module, can I use it for measuring distance between cities?

earnest forge Dec 11, 2020, 5:04 PM

#

molten hamlet i know there is some geopandas module, can I use it for measuring distance betwe...

did you plot geodata using pyplot?

dense vigil Dec 11, 2020, 5:15 PM

#

has anybody made an ai using pyttsx3, speechrecognition, wolfram alpha, and other modules?

#

if so dm me please

molten hamlet Dec 11, 2020, 5:20 PM

#

earnest forge did you plot geodata using pyplot?

???

earnest forge Dec 11, 2020, 5:21 PM

#

i just thought that you want plot the data first and then find the distance on resulted plot

#

so if it is true, you can make a function that compute the distance by its own

vocal zodiac Dec 11, 2020, 7:25 PM

#

It’s anyone there in this channel ?

lapis sequoia Dec 11, 2020, 7:26 PM

#

Hi

vocal zodiac Dec 11, 2020, 7:26 PM

#

So the project that I am referring has done the project in Google Colab, while I am trying to do the same in Jupyter Notebook, can someone give me the alternative to from google.colab import files ?

📎 image0.jpg

#

So that I can have same functionality in my jupyter notebook ?

fair kernel Dec 11, 2020, 7:30 PM

#

it runs in the browser, why u want to do this as an exe?

#

why not just dockerize or just share the link?

#

for web

#

thus i rly have no idea of how to make .exe files. so im out on that topic. but i wanted to do the same as u i guess and .exe easier to deliver to ppl sometims, so youre right

vocal zodiac Dec 11, 2020, 7:46 PM

#

So no one ?

molten hamlet Dec 11, 2020, 8:38 PM

#

earnest forge i just thought that you want plot the data first and then find the distance on r...

I got normal test.csv not related to geopandas

earnest forge Dec 11, 2020, 8:38 PM

#

did you use any specific features to build a plot

molten hamlet Dec 11, 2020, 8:39 PM

#

what are you talking about?

molten hamlet Dec 11, 2020, 8:39 PM

#

earnest forge did you use any specific features to build a plot

plot of what?

earnest forge Dec 11, 2020, 8:39 PM

#

like longtitude or latitude

molten hamlet Dec 11, 2020, 8:39 PM

#

I don;t have this data, I asked if geopandas has it

earnest forge Dec 11, 2020, 8:40 PM

#

oh

molten hamlet Dec 11, 2020, 8:40 PM

#

so I need to find data of it? 😛

#

📎 Screenshot_from_2020-12-11_21-40-41.png

#

hmm

#

wikipedia should have all citiest 🤔

molten hamlet Dec 11, 2020, 8:44 PM

#

earnest forge oh

can I create geopd dataframe from magnitute long ?

#

and just plot it ?

earnest forge Dec 11, 2020, 8:47 PM

#

do you have any geographical data in your dataset?

molten hamlet Dec 11, 2020, 8:49 PM

#

no

earnest forge Dec 11, 2020, 8:50 PM

#

hm. so how are you going to use geopandas, considering that it deals with geographical data

#

anyway, i'm not pro in it, so maybe there are some options that allow you to make more

molten hamlet Dec 11, 2020, 10:15 PM

#

earnest forge hm. so how are you going to use geopandas, considering that it deals with geogra...

https://www.kaggle.com/alexisbcook/manipulating-geospatial-data

Manipulating Geospatial Data

Explore and run machine learning code with Kaggle Notebooks | Using data from multiple data sources

#

I was looking for something like this 😄

#

but I already got data from request 🤔

dire comet Dec 12, 2020, 2:05 AM

#

Hey, anyone good at pandas that could help me? I'm not a programmer.

thin solstice Dec 12, 2020, 2:56 AM

#

would this be an appropriate channel to ask about tensorflow ?

vital yarrow Dec 12, 2020, 2:59 AM

#

what are those fancy regression error lines that make graphs look more data science-y

#

i try replicating them by plotting the regression line but adding the errors for all of my parameters and making another regression line where i subtract the errors for all parameters

#

idk

high badge Dec 12, 2020, 3:44 AM

#

📎 unknown.png

#

can someone explain to me why they use 1 and -1 instead of any other number?

#

this is for hard margin linear SVM classification

#

in the picture i believe it is a decision function with 2 inequalities or constraints

#

https://www.youtube.com/watch?v=IOetFPgsMUc&ab_channel=intriganointrigano
here is the video:

YouTube

Alexander Ihler

Support Vector Machines (1): Linear SVMs, primal form

Basics of support vector machines: definition of the margin; QP form; examples

▶ Play video

azure stump Dec 12, 2020, 6:48 AM

#

https://asr373.medium.com/finding-the-black-sheep-with-machine-learning-88c3169725ea

Medium

Finding the black sheep with Machine Learning.

Let’s find them.

tight torrent Dec 12, 2020, 7:39 AM

#

im trying to update my pypi project using twine upload dist/*

#

but it says twine is not recognized

#

its pip installed.

#

nvm i fixed it

astral path Dec 12, 2020, 7:59 AM

#

im trying to plot vertical lines over a spectogram in librosa, and i'm displaying the spectogram using librosa.display.specshow()

#

However, librosa.display doesn't have an option for plotting vertlines

#

How do I plot vertical lines placed at x values on a spectogram corresponding to a list in librosa?

#

e.g.

📎 n8amM21LyH0sgAAAABJRU5ErkJggg.png

#

cheers!

worldly palm Dec 12, 2020, 8:14 AM

#

Anyone know why animation works in Jupyter but not when run as .py?
Sample code: https://paste.pythondiscord.com/mugazuligo.yaml
(Trying to run in Pycharm IDE, but it only display the plt at the end, instead of frame by frame like in matplotlib)

📎 unknown.png

tight torrent Dec 12, 2020, 10:08 AM

#

i need help with PyPi, basically how do i update my PyPI project
whats the command
everyone is using twine upload dist/* but this dont work for me
it says twine is not recognized
;-;
i did pip install it

lapis sequoia Dec 12, 2020, 10:53 AM

#

wait a minute I thought I typed

brazen owl Dec 12, 2020, 11:42 AM

#

HI

#

can someone come on arsenic please ?

#

dat=np.loadtxt (fname=r"C:\Users\Amine13\Desktop\COURS 3I\math maintenance\a09.txt")

that the database

y=dat[:,1]
print ("Array element : ", y, "\n") 
  
a, b, c, d = stats.cumfreq(y, numbins = 4) 
  
print ("cumulative frequency : ", a) 
print ("Lower Limit : ", b) 
print ("bin size : ", c) 
print ("extra-points : ", d)```


Okay so those lines are used to get the cumulative frequencies ie like what you can see on column D.

So what i want to do is plotting f(y), and i want to do this basically:

Construire la fonction de répartition F(t) en utilisant un tableau F[30], avec F[i] = P(T ≤ ti).

i don't know how but when using those lines:
```py
t = np.linspace(0, 29, num=30)
y = np.arange(1, len(t)+1) / len(t)```

tight torrent Dec 12, 2020, 12:39 PM

#

Someone please help me with PyPi im trying to update my existing project using twine upload dist/* and its showing this error

📎 unknown.png

#

every one on internet saying thats the command but it dont work for me :/

#

please help

tight torrent Dec 12, 2020, 1:07 PM

#

help pls

charred obsidian Dec 12, 2020, 1:53 PM

#

how do I combine two dataframes? The first dataframe has a index column and the other does not. How do I add the second dataframe and auto increment the index starting with the last index of the first dataframe?

molten hamlet Dec 12, 2020, 2:11 PM

#

Hi, I got small problem, geopandas returns only one city, but there are like 10 cities with same name

#

from geopandas.tools import geocode

cities = geocode("Dobra Poland")
print(cities)

                   geometry                address
0  POINT (14.38475 53.48451)  Dobra, Poland, Poland

molten hamlet Dec 12, 2020, 2:13 PM

#

charred obsidian how do I combine two dataframes? The first dataframe has a index column and the ...

have you tried append? df1.append(df2)

charred obsidian Dec 12, 2020, 2:16 PM

#

@molten hamlet I have not. I tried pd.concat([df1, df2]), let me give it a try now.

charred obsidian Dec 12, 2020, 2:37 PM

#

@molten hamlet so apparently concat works but i had to pass in ignore_index=True

somber bane Dec 12, 2020, 2:46 PM

#

what is a suggested learning rate and number of iterations for applying sotchastics gradient descent on matrix factorization?

tight torrent Dec 12, 2020, 2:53 PM

#

my own packagae is not being found in the site-packages hence cannot be imported

#

pls help

#

📎 unknown.png

dire comet Dec 12, 2020, 2:53 PM

#

Anyone good at Pandas that can help a noobie?

tight torrent Dec 12, 2020, 2:53 PM

#

im but i cant help rn sorry

#

lost in my own prob

dire comet Dec 12, 2020, 2:54 PM

#

it's ok

#

Can I add you in case you resolve your problem?

somber bane Dec 12, 2020, 2:55 PM

#

@dire comet you can post out the problem here, may others can solve your problem too

dire comet Dec 12, 2020, 2:55 PM

#

Ok

#

Why my plot shows like this?

📎 unknown.png

#

I need it to be like this one

#

📎 unknown.png

#

I dont understand the parameters of the plots

#

Its suposed that i'm using same values

somber bane Dec 12, 2020, 2:57 PM

#

tight torrent pls help

I hope this will help https://stackoverflow.com/questions/14295680/unable-to-import-a-module-that-is-definitely-installed

Stack Overflow

Unable to import a module that is definitely installed

After installing mechanize, I don't seem to be able to import it.

I have tried installing from pip, easy_install, and via python setup.py install from this repo: https://github.com/abielr/mechaniz...

tight torrent Dec 12, 2020, 3:02 PM

#

nope

#

;-;

#

pls help

#

someone

dire comet Dec 12, 2020, 3:06 PM

#

Is there any site where I can hire a python programmer to help me with my code? is really basic

molten hamlet Dec 12, 2020, 3:07 PM

#

somber bane what is a suggested learning rate and number of iterations for applying sotchast...

smaller than 1 probably, well it depends, but 1e-3 ~ 1e-5 should be nice

molten hamlet Dec 12, 2020, 3:08 PM

#

tight torrent im but i cant help rn sorry

install it with pip, dont put it manually 😄

tight torrent Dec 12, 2020, 3:08 PM

#

.

#

.-.

#

its pip

#

only

molten hamlet Dec 12, 2020, 3:08 PM

#

google how to make setup.py
then run pip install . -e

#

and now u can edit it 😄

tight torrent Dec 12, 2020, 3:08 PM

#

bruh

#

im no noob

#

ik how packages are made

#

And that is called LOCALLY installing

molten hamlet Dec 12, 2020, 3:09 PM

#

I told you

#

google setup.py

tight torrent Dec 12, 2020, 3:09 PM

#

im trying to install from pypi

molten hamlet Dec 12, 2020, 3:09 PM

#

pypi uses pip

tight torrent Dec 12, 2020, 3:10 PM

#

thats what im using

molten hamlet Dec 12, 2020, 3:11 PM

#

pypi has always command for instaling

tight torrent Dec 12, 2020, 3:50 PM

#

Guys i need to create Enviorment Variables for my api keys but where do i make it
its running on pypi
i tried on desktop but it no worke..

molten hamlet Dec 12, 2020, 4:20 PM

#

Did anybody used fuzzywuzzy?
I wonder If I can reduce 170 labesl into few

lapis sequoia Dec 12, 2020, 4:42 PM

#

can somebody help me with something related to matplotlib?

#

disregard

dire comet Dec 12, 2020, 4:58 PM

#

Anyone know how to make a matrix A vs B, from a DataFrame?

molten hamlet Dec 12, 2020, 6:17 PM

#

lapis sequoia can somebody help me with something related to matplotlib?

ask question

molten hamlet Dec 12, 2020, 6:17 PM

#

dire comet Anyone know how to make a matrix A vs B, from a DataFrame?

ask question

molten hamlet Dec 12, 2020, 6:35 PM

#

🥂

lapis sequoia Dec 12, 2020, 6:55 PM

#

Is there a recommended tutorial for Deep Learning in python?

molten hamlet Dec 12, 2020, 6:57 PM

#

lapis sequoia Is there a recommended tutorial for Deep Learning in python?

no, but check pins

lapis sequoia Dec 12, 2020, 6:58 PM

#

Thanks!

sterile minnow Dec 12, 2020, 7:11 PM

#

😄

lunar bear Dec 12, 2020, 8:21 PM

#

I wanna do data science where should I start.

blissful merlin Dec 12, 2020, 8:38 PM

#

I just posted in #help-cheese about a neural network problem I having - in particular to do with tensorflow input shape.

if there are any experts would hugely appreciate their help :)

arctic wedgeBOT Dec 12, 2020, 8:46 PM

#

:x: According to my records, this user already has a mute infraction. See infraction #24333.

#

:incoming_envelope: :ok_hand: applied mute to @lapis sequoia until 2020-12-12 20:56 (9 minutes and 59 seconds) (reason: burst rule: sent 9 messages in 10s).

astral path Dec 12, 2020, 10:02 PM

#

im trying to plot vertical lines over a spectogram in librosa, and i'm displaying the spectogram using librosa.display.specshow()
However, librosa.display doesn't have an option for plotting vertlines
How do I plot vertical lines placed at x values on a spectogram corresponding to a list in librosa?
e.g.

📎 n8amM21LyH0sgAAAABJRU5ErkJggg.png

dire comet Dec 13, 2020, 12:06 AM

#

Anyone knows what this mean?
KeyError: 139028

bitter harbor Dec 13, 2020, 12:09 AM

#

In what context?

idle yoke Dec 13, 2020, 12:12 AM

#

dire comet Is there any site where I can hire a python programmer to help me with my code? ...

I would suggest fiverr

dire comet Dec 13, 2020, 12:13 AM

#

bitter harbor In what context?

I was making a Matrix from a DataFrame

#

And used this

#

val = df.loc[i,"Sector"]

#

in a For in

#

Forget it, found the error. I used In the for in Range (0,len(df) +1) and was without the +1

weak solstice Dec 13, 2020, 12:44 AM

#

does anyone know how to set up gpu acceleration for tensorflow

midnight crag Dec 13, 2020, 12:45 AM

#

hello nix

void onyx Dec 13, 2020, 1:01 AM

#

im here to learn how to code and make bots

lilac oasis Dec 13, 2020, 1:44 AM

#

Anyone able to help with streamlit and creating different charts based on data from csv?

tight torrent Dec 13, 2020, 4:09 AM

#

Guys my environ variable is printing nOne
when im trying to print it
even though its registered in my system

lapis sequoia Dec 13, 2020, 6:09 AM

#

Who is interests by lot of knowledge ?

molten hamlet Dec 13, 2020, 9:26 AM

#

lapis sequoia Who is interests by lot of knowledge ?

me

#

feed me

torpid cave Dec 13, 2020, 10:45 AM

#

Sup

#

who is familiar with SeetingWithCopyWarning.. a vlue is trying to be set on a copy of a slice from a DataFrame

hushed wasp Dec 13, 2020, 3:19 PM

#

Hello guys,

I would like to represent my errors (y - ypred) after a pipeline.
Which is the easiest way please?

Thanks

lapis sequoia Dec 13, 2020, 4:13 PM

#

can someone help with a big dataset, i know how to code it but it is taking me a really long time for my code to run

molten hamlet Dec 13, 2020, 5:18 PM

#

what was the name of method to give more data to the model to prevent overfitig?

#

regularization

#

thanjks

dapper pivot Dec 13, 2020, 6:03 PM

#

Hi there, a BI developer here interested in applied ML, econometrics and how to use applied data science techniques in predictive analytics!

upper schooner Dec 13, 2020, 7:06 PM

#

Hi im trying to use images from an HDR movie (10bit depth) to train on, but i cant find a single package that can load my video without compressing the video down to 8bit. I've heard that ffmpeg can do it, but ive tried and failed, and the documentation is so confusing i don't know where to even start reading. If anyone has experience with how to do this i would appreciate the help.

upper schooner Dec 13, 2020, 8:17 PM

#

lapis sequoia can someone help with a big dataset, i know how to code it but it is taking me a...

Look into data loaders and generators if youre not already using them. This article explains it really well: https://stanford.edu/~shervine/blog/pytorch-how-to-generate-data-parallel

A detailed example of data loaders with PyTorch

Blog of Shervine Amidi, Graduate Student at Stanford University.

burnt veldt Dec 14, 2020, 12:21 AM

#

Why cant I load a excel file with pandas in jupiter notebook its a xlsx file type

#

i know i have to use openpyxl

#

But even with that it sitll doesnt work

#

df1 = pandas.read_excel("supermarkets.xlsx", sheet_name=0, engine = openpyxl)
df1

#

which raises the error

#

ValueError: Unknown engine: <module 'openpyxl' from '/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/site-packages/openpyxl/init.py'>

lapis sequoia Dec 14, 2020, 12:35 AM

#

What's data science

molten hamlet Dec 14, 2020, 1:49 AM

#

lapis sequoia What's data science

data mining is old term

serene scaffold Dec 14, 2020, 4:18 AM

#

lapis sequoia What's data science

the science of data

#

or more specifically, how you can find insights from large amounts of data by leveraging computer science.

#

It's closely related to AI and machine learning.

charred cipher Dec 14, 2020, 8:11 AM

#

hi so ive been working on this project and the first sight errors are gone but slowly all the inner errors are coming in that come only when you execute and try

#

ill paste bin the program for you all to see and try

#

since ive imported a csv file from my files, idk if you all will be able to run this program but do let me know if you spot some errors in places like histogram, pie chart etc.

#

https://paste.pythondiscord.com/tonomorose.py

#

ive paste the code here

molten hamlet Dec 14, 2020, 10:31 AM

#

no database mininig dude

#

data mining

#

Data mining, also called knowledge discovery in databases, in computer science, the process of discovering interesting and useful patterns and relationships in large volumes of data. The field combines tools from statistics and artificial intelligence (such as neural networks and machine learning) with database management to analyze large digital collections, known as data sets. Data mining is widely used in business (insurance, banking, retail), science research (astronomy, medicine), and government security (detection of criminals and terrorists).

#

britanica

molten hamlet Dec 14, 2020, 11:18 AM

#

it started with data mining

#

now its data science

#

is it too hard to understand ?

torpid cave Dec 14, 2020, 11:34 AM

#

Hi all

#

I have a pandas question for which I think I am having more problem than expected

molten hamlet Dec 14, 2020, 11:41 AM

#

ahh, I just meant its older then now 😄

#

like parent

rich plover Dec 14, 2020, 12:01 PM

#

Hi where is a good place, to see the analysis of a given data set

polar charm Dec 14, 2020, 1:23 PM

#

Hello, I'm getting py UndefinedMetricWarning: R^2 score is not well-defined with less than two samples. and very fluctuating results on my predictions.

#

49.49375681 this time and 1297.54128991 the next time. How do I fix it

bold olive Dec 14, 2020, 1:38 PM

#

How can I use the sk-learn classifiers using just a single feature?

sweet plaza Dec 14, 2020, 2:05 PM

#

Does anyone have a good book or youtuibe video that teaches really baseline Data Literacy , I think its jsut hard to study Data Science without knowing fundamentals like these .

Much appreciated

frozen skiff Dec 14, 2020, 3:02 PM

#

polar charm Hello, I'm getting ```py UndefinedMetricWarning: R^2 score is not well-defined w...

Regarding the error/warning: you seem to be calculating R^2 on less than two samples, which results in division by zero. Consider the formula in https://scikit-learn.org/stable/modules/model_evaluation.html#r2-score if you take n = 1.

What happens is that the average y̅ is equal to yᵢ, resulting in the denominator being zero and causing a divison by zero

scarlet storm Dec 14, 2020, 3:03 PM

#

can I ask web scrapping question here ?

polar charm Dec 14, 2020, 3:03 PM

#

So how do I get more accuracy on my predictions? and is it ```py
r2_score
X = ['Adj Close']
y = ['Adj Close']



Edit: did  ```py
r2_Score = X = df['Adj Close'], y = df['Adj Close'] ```
got error ```py
r2_Score = X = df['Adj Close'], y = df['Adj Close']
ValueError: too many values to unpack (expected 2) ```

frozen skiff Dec 14, 2020, 3:11 PM

#

How do i get more accuracy on my predictions?
This really depends on what problem you are solving, but a very general tip would be to just check if somebody else has done something on your specific topic of interest

polar charm Dec 14, 2020, 3:12 PM

#

I'm trying to code a s&p500 predictor.

scarlet storm Dec 14, 2020, 3:16 PM

#

I am trying to scrape genres and reviews out of imdb page and I am able to scrape reviews but not genre cause I think js is loading it dynamically

#

what can I do to scrape the genre in this case ?

frozen skiff Dec 14, 2020, 3:21 PM

#

scarlet storm I am trying to scrape genres and reviews out of imdb page and I am able to scrap...

If the website is dynamic, you will have to use Selenium or some alternative

scarlet storm Dec 14, 2020, 3:22 PM

#

thanks

flint nest Dec 14, 2020, 4:02 PM

#

why is a conv2d 2d kernel used to convolve over rgb images?

#

in tensorflow

#

like how does it work, you would expect a 3d kernel for rgb images right?

frozen moth Dec 14, 2020, 4:15 PM

#

Hey guys, quick question. Say I have a dictionary:

my_dict = {'male': ['Jack', 'John', 'Bob'], 'female': ['Jane', 'Mary', 'Carolina']}

and I also have a DataFrame with a feature filled with names:

my_data = {'id' : [1, 2, 3], 'name' : ['John', 'Mary', 'Jack']}
my_dataframe = pd.DataFrame(data = my_data)

and I want to create a new feature called gender:

my_data = {'id' : [1, 2, 3], 'name' : ['John', 'Mary', 'Jack'], 'gender' : ['male', 'female', 'male']}
my_dataframe = pd.DataFrame(data = my_data)

How can I create my for loop such that for every entry of my DataFrame male or female gets added to the new column based on the names that are in my dictionary and weather or not they match the key-item pairs?

frozen skiff Dec 14, 2020, 4:39 PM

#

you could iterate over names ("name") inside your my_data dictionary, check whether (1) the name is inside the male names' list or (2) is inside female names' list and store that inside a new list, then assign it to a new column inside your dataframe

steep widget Dec 14, 2020, 6:06 PM

#

hi guys, this is probably answered but i couldn't find a good way to word it on google.
say i have a list with elements structured like ['1000 +xx']
where xx is some uncertainty that changes with each element, e.g [ ['1000 +44], ['2313 +54] ]
how to i remove the +xx from each element, something like cut everything starting from +?

thank you 🙇‍♂️

muted oyster Dec 14, 2020, 6:10 PM

#

can someone help me with interpreting the result of Anova test ? like if I have to find significance between Age and Sales, and the p value is coming below Alpha(0.05)

frozen skiff Dec 14, 2020, 6:23 PM

#

steep widget hi guys, this is probably answered but i couldn't find a good way to word it on ...

If I understand correctly that you're dealing with strings, you could take a look at the find() function to find the starting index of "+", and then slice your string accordingly

lapis sequoia Dec 14, 2020, 7:02 PM

#

@ me

slim tiger Dec 14, 2020, 7:05 PM

#

marker="."

tepid pawn Dec 14, 2020, 7:07 PM

#

Hey everyone, new to Discord. This place is loaded with great info and experts.

teal sluice Dec 14, 2020, 7:29 PM

#

I have a dataframe loaded with one column being dates, what be the best way of creating a dataframe specific to the dates for the month/year (the format of the date is year-month-day all as integers)??

lapis sequoia Dec 14, 2020, 8:25 PM

#

any good courses to get used to working with data? @ me

heady hatch Dec 14, 2020, 8:28 PM

#

If you’re using pandas, you can strip the dates using something like “df.dt.month”. They need to be in a date time readable format.

#

@teal sluice

tepid pawn Dec 14, 2020, 8:35 PM

#

@lapis sequoia The datacamp ones seemed to be ok. I had a free subscription and we had to do a couple of them with a class I was taking. To be honest all of the ones that I've looked at online have been very regurgitative in nature. If you know Python (or R) I'd recommend finding a project on Kaggle or wherever and just go at it. You'll screw it up at first, but as you work through it you'll learn. I google the hell out of stuff that I don't know and watch different videos online for specific things.

lapis sequoia Dec 14, 2020, 8:36 PM

#

thanks man!

tepid pawn Dec 14, 2020, 8:37 PM

#

You're welcome

languid dagger Dec 15, 2020, 12:50 AM

#

Is there a faster way to turn a 2d numpy array into a 1d numpy array of tuples than numpy.array(list(map(tuple, my_2d_nparray)))? And the followup if anyone is willing to engage, is whether I'm barking up the wrong tree entirely.

heady hatch Dec 15, 2020, 12:59 AM

#

languid dagger Is there a faster way to turn a 2d numpy array into a 1d numpy array of tuples t...

As in you want to flatten the arrays?

languid dagger Dec 15, 2020, 1:05 AM

#

The context is that I'm saving a point cloud as a PLY file. I have an array of XYZ coordinate and RGBA color data to pass to plyfile, but it complains if it has more than one dimension.

#

https://github.com/dranjan/python-plyfile

GitHub

dranjan/python-plyfile

NumPy-based text/binary PLY file reader/writer for Python - dranjan/python-plyfile

tepid pawn Dec 15, 2020, 3:25 AM

#

Anyone on here tonight?

serene scaffold Dec 15, 2020, 5:16 AM

#

tepid pawn Anyone on here tonight?

yes

tepid pawn Dec 15, 2020, 5:45 AM

#

I'm new here, just trying to get a feel for it.

#

working through a problem set for predictions right now

lapis sequoia Dec 15, 2020, 9:47 AM

#

hey guys

#

I need some urgent help in pandas and matplotlib

#

can someone help me in the DMs?

compact flume Dec 15, 2020, 9:59 AM

#

Hey y'all! I'm still kinda new to python but was wondering if there are any beginner friendly data science/quantitative modeling project ideas I could start on?

velvet thorn Dec 15, 2020, 10:11 AM

#

steep widget hi guys, this is probably answered but i couldn't find a good way to word it on ...

regular expressions

velvet thorn Dec 15, 2020, 10:11 AM

#

frozen moth Hey guys, quick question. Say I have a dictionary: ```Python my_dict = {'male': ...

are there cases where names are in neither?

velvet thorn Dec 15, 2020, 10:11 AM

#

flint nest like how does it work, you would expect a 3d kernel for rgb images right?

why do you say so?

#

because images are 3D?

carmine bough Dec 15, 2020, 10:22 AM

#

Hi guys, I need help with the following: I have a Video in which a person lifts her right arm, standing in different positions. I also have pictures, that show exactly this pose. Now I need a programme, that uses machine learning, saying when the pose "right arm up" is shown in the video. I already tried to find something like that but counldn't find anything. Is there any github repository or so that provides code doing exactly this? Or is someone able to write that code for me in return for a little fee?

lapis sequoia Dec 15, 2020, 10:43 AM

#

carmine bough Hi guys, I need help with the following: I have a Video in which a person lifts ...

Hi, why don't you check for something like gesture recognition then you can relate that to your project. Just a simple google search gave me this https://github.com/SouravJohar//handy I think it's similar to what you are trying to do!

GitHub

SouravJohar/handy

An easy to use wrapper (OpenCV) for hand recognition - SouravJohar/handy

carmine bough Dec 15, 2020, 11:58 AM

#

lapis sequoia Hi, why don't you check for something like gesture recognition then you can rela...

Thank you, thats not too bad, but not using machine learning. However this has to be a part of the project...

lapis sequoia Dec 15, 2020, 12:00 PM

#

I think they are using machine learning, to determine what the gesture is, it's not possible to just do it with opencv

velvet thorn Dec 15, 2020, 12:16 PM

#

carmine bough Thank you, thats not too bad, but not using machine learning. However this has t...

...is this for school?

upbeat jetty Dec 15, 2020, 12:23 PM

#

Which free (or at least having a free tier) APIs you think are most useful to learn?

fleet plover Dec 15, 2020, 2:42 PM

#

For https://gist.github.com/promach/a68da1a80e84a55320e73501711618c8#file-mcts-py-L41 , how should I deal with parent_visits inside MCTS ?

Gist

A simple vanilla Monte-Carlo Tree Search implementation in python

A simple vanilla Monte-Carlo Tree Search implementation in python - mcts.py

safe tapir Dec 15, 2020, 3:51 PM

#

Is there a jargon etymology resource? Trying to understand why certain words were used to describe things so I can more easily remember them. Some things are google-able (latent -> hidden), but some are not as obvious (e.g. why are they called transformers?)

hushed wasp Dec 15, 2020, 4:41 PM

#

Hello Guys,

Does someone could tell me why I have a worse score after a GridSearch than before?
On a KNRegressor
Here is mycode :

#

📎 unknown.png

lapis iris Dec 15, 2020, 4:58 PM

#

Hi I want to remove 3rd elements from nested list

lista = [('one', 'two', ['remove this element']),
         ('three', 'four', ['remove this element'])]

any advice how to do this shortly without iterate over and over?

heady hatch Dec 15, 2020, 4:59 PM

#

hushed wasp Hello Guys, Does someone could tell me why I have a worse score after a GridSea...

One score is referring to r2 and the other one is referring to mean accuracy. How do you know it has a worse score?

heady hatch Dec 15, 2020, 5:00 PM

#

lapis iris Hi I want to remove 3rd elements from nested list ```Python lista = [('one', 'tw...

Is it all in one list?

lapis iris Dec 15, 2020, 5:02 PM

#

heady hatch Is it all in one list?

kindof, list contains around 50 records

heady hatch Dec 15, 2020, 5:03 PM

#

Like 50 records under one list?

lapis iris Dec 15, 2020, 5:03 PM

#

ye

heady hatch Dec 15, 2020, 5:04 PM

#

I don’t think you need to iterate through anything at all.

If you want to delete every nth element from a list you can just use

del a[n-1::n]

#

Oh wait to clarify do you want to keep the removed elements?

lapis iris Dec 15, 2020, 5:05 PM

#

elements are hardcoded si I dont care bout em

heady hatch Dec 15, 2020, 5:06 PM

#

Then yea just delete every nth starting from n-1.

lapis iris Dec 15, 2020, 5:08 PM

#

heady hatch Then yea just delete every nth starting from n-1.

well,

print(lista[:-1])

doesnt work

heady hatch Dec 15, 2020, 5:10 PM

#

lapis iris well, ``` print(lista[:-1]) ``` doesnt work

What about it doesn't work?

#

What are you expecting print(lista[:-1]) to do?

lapis iris Dec 15, 2020, 5:11 PM

#

output should be like [('one', two'), ('three', 'four')]
and output is like [('one', 'two', ['remove this element'])]

heady hatch Dec 15, 2020, 5:12 PM

#

Oh I think I misunderstood your list.

#

It's a list of tuples?

lapis iris Dec 15, 2020, 5:12 PM

#

ye

heady hatch Dec 15, 2020, 5:12 PM

#

Then yea you're going to have to iterate.

#

Unless you want to flatten and zip it back up again.

#

new_a = [e[:2] for e in a]

#

Something like that should work.

hushed wasp Dec 15, 2020, 5:14 PM

#

heady hatch One score is referring to r2 and the other one is referring to mean accuracy. Ho...

For me my R2 went to 0.70 to 0.48

heady hatch Dec 15, 2020, 5:14 PM

#

hushed wasp For me my R2 went to 0.70 to 0.48

Why do you think that 0.48 is r2?

lapis iris Dec 15, 2020, 5:15 PM

#

I went with: print([i[:2] for i in lista])

hushed wasp Dec 15, 2020, 5:15 PM

#

heady hatch Why do you think that 0.48 is r2?

The KNRegressor documentary

heady hatch Dec 15, 2020, 5:15 PM

#

Ahh yea I totally misread and thought it was a classifier. my bad.

heady hatch Dec 15, 2020, 5:18 PM

#

hushed wasp The KNRegressor documentary

Do you mind doing cross validation on your original model? I'm trying to find some more insight about the data.

#

I'm wondering if there's a distribution difference or if the gridsearch is tuned towards some other goals than what you originally wanted.

hushed wasp Dec 15, 2020, 5:20 PM

#

I guess it's something like that

#

I gonna try it

#

my original model is like an other one?
I don't have any trouble with RF or ridge or lasso only with the KN...

heady hatch Dec 15, 2020, 5:32 PM

#

Oh as in the cv from original model match cv from gridsearch?

languid dagger Dec 15, 2020, 5:45 PM

#

I'm trying to segment 2D points according to whether they "form a line". Like whether the local dimension is 1 or 2. I've had some imperfect success by dilating then eroding, which removes the lines and keeps the blobs. It seems like a problem that might already be solved much better though. Anybody know?

📎 unknown.png

hushed wasp Dec 15, 2020, 5:53 PM

#

heady hatch Oh as in the cv from original model match cv from gridsearch?

I guess... Sorry I am still very new in all of this

plain cargo Dec 15, 2020, 8:25 PM

#

hi all. I am new to pandas and dataframes. Apparently dataframes damage the type system as it does not reveals its columns/field names until runtime. But converting between dataframes and list of model objects seems too cumbersome. Is there an elegant way to solve this issue?

heady hatch Dec 15, 2020, 9:08 PM

#

hushed wasp I guess... Sorry I am still very new in all of this

Heya no worries. Your situation is pretty interesting too. I'm trying to think of reasons why gridsearch might find a lower optimal point than default setting.

One basic thing you can do is add the default parameters in there too and see if gridsearch finds that. If not then I believe something is really wrong.

teal sluice Dec 15, 2020, 9:09 PM

#

I have a dataframe (imported via csv) which consists of 2 columns, one being only 2 values ( Outgoing/Ingoing) and another consisting of an amount, I was wondering if there was a way of changing the secnd column so all the rows consisting of 'Outgoing' has a '-' in front of the amount??

heady hatch Dec 15, 2020, 9:17 PM

#

teal sluice I have a dataframe (imported via csv) which consists of 2 columns, one being onl...

Heya,

to clarify your dataframe has two columns, one (in/out) and the other amount?

#

Assuming you're working with pandas.

You can boolean index the outgoing and either multiply the column by -1 or add '-' in front if it's a string.

#

ie.

df 
  a   b
  in  3
  out 4
  in  2
  out 5

mask = df['a'] == 'out'
neg_col = df.loc[mask, 'b'] * -1
df.loc[mask, 'b'] = neg_col

#

You can clean that all up in one line if you'd like. I separated it for better readability.

hushed wasp Dec 15, 2020, 9:23 PM

#

heady hatch Heya no worries. Your situation is pretty interesting too. I'm trying to think o...

I found some others cases on stackoverflow with the KNRegressor where there is the same issue.
Can I ask you to explain me how to do a default gridsearch please? 🙂

heady hatch Dec 15, 2020, 9:33 PM

#

Oh by default gridsearch, I just meant using the hyperparameter you used in the original model where you just fit it on the xtrain and ytrain.

But looking over your parameter dictionary, I think I see the default hyperparameters there.

Can you try running gridsearch with nothing but 5 neighbors, uniform weights, auto algorithm, and minkowski metric?

#

@hushed wasp

#

See if gridsearch comes up with the same model as just the regular model.

#

oh there's probably a scoring section too in gridsearchcv.

#

In terms of scoring, you can put 'r2'.

#

So like

#

gridsearchcv(model, parameter, scoring='r2')

teal sluice Dec 15, 2020, 9:48 PM

#

@heady hatch Works perfectly thnx, do u mind just explaining what the code does line by line, just so I understand what's happening 😄

heady hatch Dec 15, 2020, 9:48 PM

#

teal sluice <@!542872811245666305> Works perfectly thnx, do u mind just explaining what the ...

Yea of course.

To preface, how familiar are you with pandas?

teal sluice Dec 15, 2020, 9:49 PM

#

I just know the basics essentially, so basic selecting, indexing etc

heady hatch Dec 15, 2020, 9:50 PM

#

Ahh okay okay. Are you familiar with boolean indexing yet?

teal sluice Dec 15, 2020, 9:51 PM

#

not really

heady hatch Dec 15, 2020, 9:51 PM

#

If you don't mind, on the side can you pull up a pandas dataframe so you can follow along.

teal sluice Dec 15, 2020, 9:51 PM

#

Yep will do

heady hatch Dec 15, 2020, 9:52 PM

#

here's a quick df you can initialize.

df = pd.DataFrame({'a': list('abcd'), 'b': list('efgh')})

#

You should have a dataframe of shape 4 x 2, where both columns are strings.

teal sluice Dec 15, 2020, 9:54 PM

#

yep just put it onto jupyter

heady hatch Dec 15, 2020, 9:54 PM

#

Nice nice.

#

So now print this.

print(df['a'])

and let me know what you see.

teal sluice Dec 15, 2020, 9:55 PM

#

prints a column a out with the index on the left

earnest forge Dec 15, 2020, 9:58 PM

#

Have anyone tried to implement Linear Regression by themselves? Without using sklearn. Just in educational purposes to understand principles of model work much better

heady hatch Dec 15, 2020, 9:59 PM

#

teal sluice prints a column a out with the index on the left

Okay now try printing

print(df['a'] == 'c')

Let me know what you're seeing now.

heady hatch Dec 15, 2020, 9:59 PM

#

earnest forge Have anyone tried to implement Linear Regression by themselves? Without using sk...

Do you have any question on implementation itself?

teal sluice Dec 15, 2020, 9:59 PM

#

prints out the index with a list of boolean values

#

guessing it goes through and sees which is c

#

and thus marks that one as true

heady hatch Dec 15, 2020, 10:00 PM

#

Yup.

and now try printing

print(df[df['a'] == 'c'])

teal sluice Dec 15, 2020, 10:02 PM

#

prints out the one row which has c in column a

heady hatch Dec 15, 2020, 10:05 PM

#

Yup, and this is boolean indexing.

#

Essentially what it does is it selects indices by boolean values.

teal sluice Dec 15, 2020, 10:06 PM

#

ohh ok

#

so from the code u suggested, it select all the rows which are 'out' and to those it multiplies whatever is in the second columns by -1 to give the negative of that answer

heady hatch Dec 15, 2020, 10:07 PM

#

Now to explain what I've written.

mask = df['a'] == 'out' #get indices of rows with Output
neg_col = df.loc[mask, 'b'] * -1 #select column b from all those rows and multiply it by -1
df.loc[mask, 'b'] = neg_col #reassign those rows back into the original dataframe

#

Mhm.

teal sluice Dec 15, 2020, 10:08 PM

#

thanks for the help really do appreciate it

heady hatch Dec 15, 2020, 10:08 PM

#

Technically you can do it all in one line.

like

df.loc[df['a'] == 'out', 'b'] = df.loc[df['a'] == 'out', 'b'] * -1

But there's a lot of things going on here.

#

No problem, glad to be of help.

#

Or maybe you can just

df.loc[df['a'] == 'out', 'b'] *= -1

But I've never tried it.

bitter harbor Dec 15, 2020, 11:46 PM

#

i'd assume it'd be pretty standard, Ive used it on pycharm but it's nothing special

#

vsc might have an actual linter extension tho

rugged flame Dec 16, 2020, 12:25 AM

#

📎 image0.png

#

Can anybody help guys?

lapis sequoia Dec 16, 2020, 1:23 AM

#

So I am using this software https://augcog.github.io/ROAR/ in this autonomous driving simulation game http://carla.org

#

I have this super annoying issue where the car stays inbetween the lanes when it drives straight

#

But as soon as the road curves

#

my program gets confused? Like it can’t see the lines properly and it keeps crossing them

#

It’s like it is driving between the curved lanes as if they are straight

#

if anyone has experience working with Carla please @ me

twilit pilot Dec 16, 2020, 1:47 AM

#

When I am using numpy, I want to append and array to another array and get a 2d array, but this is what I am getting. Can someone help?

import numpy as np

final = np.array([])
final = np.append(np.array([1, 2, 3]), final)
final = np.append(np.array([4, 5, 6]), final)
final = np.append(np.array([7, 8, 9]), final)

print(final)
```Result :

[7. 8. 9. 4. 5. 6. 1. 2. 3.]
But I want it to be
[[1, 2, 3], [4, 5, 6], [7, 8, 9]]

heady hatch Dec 16, 2020, 1:58 AM

#

twilit pilot When I am using numpy, I want to append and array to another array and get a 2d ...

You can try

l = []
l.append([1, 2, 3])

np.array(l)

bitter harbor Dec 16, 2020, 2:32 AM

#

concatenate lets you choose the axis as well

velvet thorn Dec 16, 2020, 2:37 AM

#

twilit pilot When I am using numpy, I want to append and array to another array and get a 2d ...

you want r_

#

or stack

#

!e

import numpy as np

a1 = [1, 2, 3]
a2 = [4, 5, 6]
a3 = [7, 8, 9]

print(np.stack([a1, a2, a3]))

arctic wedgeBOT Dec 16, 2020, 2:38 AM

#

@velvet thorn :white_check_mark: Your eval job has completed with return code 0.

001 | [[1 2 3]
002 |  [4 5 6]
003 |  [7 8 9]]

vale rampart Dec 16, 2020, 2:40 AM

#

hi

twilit pilot Dec 16, 2020, 3:05 AM

#

velvet thorn !e ```py import numpy as np a1 = [1, 2, 3] a2 = [4, 5, 6] a3 = [7, 8, 9] print...

ah ok thanks man

earnest forge Dec 16, 2020, 7:03 AM

#

heady hatch Do you have any question on implementation itself?

No

livid quartz Dec 16, 2020, 9:16 AM

#

Does anyone know what cmap this is? I'm using matplotlib

📎 unknown.png

gray tartan Dec 16, 2020, 10:02 AM

#

@livid quartz it looks like bwr

#

here's the reference for all cmaps matplotlib has https://matplotlib.org/examples/color/colormaps_reference.html

#

btw, if someone can look at it, i'm having a problem with matplotlib subplots, im using a gridspec and shared axis for a grid of subplots, this way :

fig = plt.figure(figsize=(np.array(img.shape) / 20)[1::-1])
gs = fig.add_gridspec(5, 5, wspace=0, hspace=0)
plt.xticks(np.arange(5))
plt.yticks(np.arange(5), np.arange(1, 6))

problem is, i want the x axis ticks and labels to be at the top, so i added that to my setup :

plt.rcParams['xtick.bottom'] = False
plt.rcParams['xtick.labelbottom'] = False
plt.rcParams['xtick.top'] = True
plt.rcParams['xtick.labeltop'] = True

but only the ticks go to the top, the label stay at the bottom, do you have any idea of why ?

lapis sequoia Dec 16, 2020, 10:15 AM

#

I envy you guys

#

I’ve been studying my ass off

#

Still barely know anything

livid quartz Dec 16, 2020, 10:30 AM

#

gray tartan here's the reference for all cmaps matplotlib has https://matplotlib.org/example...

thanks

hushed wasp Dec 16, 2020, 10:59 AM

#

heady hatch Oh by default gridsearch, I just meant using the hyperparameter you used in the ...

Thanks a lot gonna try it 🙂

torpid wadi Dec 16, 2020, 11:02 AM

#

in pandas I need to filter column that contains list when I try

filtered_data.query('"a" in letters')

letters is ['a', 'b'] I get error SystemError
I found documentation only for oposite
df[df['b'].isin(["a", "b", "c"])]

main badger Dec 16, 2020, 11:30 AM

#

Hi.. I have a nested dictionary I want to plot. With the below code, I get the plot grouped by the inner dictionary key on the x-axis. I want the plot to be grouped by the outer dictionary key. What do I do?

df = pd.DataFrame(precision_data)
df.plot(kind="bar", stacked=False)
plt.show()

main badger Dec 16, 2020, 12:00 PM

#

Found it. transpose() does it.

df.transpose().plot(kind="bar", stacked=False)

lapis sequoia Dec 16, 2020, 12:20 PM

#

Hi all I keep getting the error there is no file.directory. I think Python is looking in the wrong directory but I can't seem to figure it out:

import os
import pandas as pd
path = 'D:\Python\FileForMergeExcel'
cwd = os.path.abspath(path)
files = os.listdir(cwd)

df = pd.DataFrame()
for file in files:
    if file.endswith('.xlsx'):
        df = df.append(pd.read_excel(file), ignore_index=True)
df.head()
df.to_excel('total.xlsx') `` 

``FileNotFoundError: [Errno 2] No such file or directory: 'Report 1 - Copy.xlsx```

livid quartz Dec 16, 2020, 12:29 PM

#

why does this give me the column means? I thought axis 0 indicates rows?

📎 unknown.png

gray tartan Dec 16, 2020, 12:34 PM

#

@livid quartz it averages along the axis, it's like putting your data in a table and adding a row which has the mean of every column

livid quartz Dec 16, 2020, 12:35 PM

#

Ah, that explains it thank you

gray tartan Dec 16, 2020, 12:36 PM

#

@lapis sequoia you need to escape the \ because they ain't interpreted as characters

lapis sequoia Dec 16, 2020, 12:37 PM

#

Thanks man where to put the \ ?

gray tartan Dec 16, 2020, 12:37 PM

#

path = 'D:\\Python\\FileForMergeExcel'

lapis sequoia Dec 16, 2020, 12:37 PM

#

You're the best Imma try it out

gray tartan Dec 16, 2020, 12:37 PM

#

i would use pathlib for that tho, it handles paths better

lapis sequoia Dec 16, 2020, 12:42 PM

#

@gray tartan I think it's an other problem because it is finding the files but still gives the error. I will try pathlib

📎 GetAttachmentThumbnail.png

lapis sequoia Dec 16, 2020, 1:06 PM

#

Still the same error :C

hushed wasp Dec 16, 2020, 1:17 PM

#

hushed wasp Thanks a lot gonna try it 🙂

@heady hatch it works with a gridsearch with normal parameters!!

torpid cave Dec 16, 2020, 1:53 PM

#

@lapis sequoia try using /

gray tartan Dec 16, 2020, 2:29 PM

#

@lapis sequoia sry i'm working, well windows uses \ in their paths, so it should work, the problem might come from the rest of your path then

lapis sequoia Dec 16, 2020, 2:45 PM

#

The weird thing is that the files are in the directory that I use

📎 GetAttachmentThumbnail.png

frozen moth Dec 16, 2020, 4:00 PM

#

velvet thorn are there cases where names are in neither?

nope, but I figured it out in the end using lambda expressions

#

but thanks guys

#

Guys easy and quick question:
If i have a dataFrame with duplicated names and with different values how to I aggregate them into one single entry?

df = pd.DataFrame(
    data = [
        ['John', 1, 1],
        ['Tom', 0, 1],
        ['Mary', 1, 4],
        ['Tom', 3, 1],
        ['John', 0, 3]],
    columns = ['Name', 'Dogs_Owned', 'Cats_Owned']
)

#

from this

📎 Screenshot_2020-12-16_at_15.57.15.png

#

to this

📎 Screenshot_2020-12-16_at_15.57.20.png

#

figured it out

#

solved

storm gate Dec 16, 2020, 4:26 PM

#

so are more people using fastAPI over flask now?? at least for model deployment?

#

Ive seen a few things where the code looks identical between the two

shy moat Dec 16, 2020, 4:41 PM

#

Can I configure the depth of profiling?
I want to analyze computation time of my own code but the default profiling is too detailed.

dapper karma Dec 16, 2020, 4:47 PM

#

is there a library that can locate an image on the screen quickly? The pyautogui version of this function (pyautogui.locateCenterOnScreen) takes around 1 or 2 seconds so it's pretty slow

storm gate Dec 16, 2020, 4:48 PM

#

opencv

#

they have a function (blanking on what its called) where u provide a sample image and it tries to find it or all instances of it on the screen

#

https://docs.opencv.org/master/d4/dc6/tutorial_py_template_matching.html

dapper karma Dec 16, 2020, 4:49 PM

#

alright ill check that out

#

hmm it doesnt say how fast that is

#

i think i'll manually test to see if its faster than pyautogui

storm gate Dec 16, 2020, 4:55 PM

#

its def gonna be faster then pyautogui

rich plover Dec 16, 2020, 4:55 PM

#

X = df[["xG","deep", "ppda_coef"]].values.reshape(-1,1)
Y = df[["scored"]].values

#

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-8-96914964615e> in <module>
----> 1 X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.2, random_state=7)

~\anaconda3\lib\site-packages\sklearn\model_selection\_split.py in train_test_split(*arrays, **options)
   2125         raise TypeError("Invalid parameters passed: %s" % str(options))
   2126 
-> 2127     arrays = indexable(*arrays)
   2128 
   2129     n_samples = _num_samples(arrays[0])

~\anaconda3\lib\site-packages\sklearn\utils\validation.py in indexable(*iterables)
    291     """
    292     result = [_make_indexable(X) for X in iterables]
--> 293     check_consistent_length(*result)
    294     return result
    295 

~\anaconda3\lib\site-packages\sklearn\utils\validation.py in check_consistent_length(*arrays)
    254     uniques = np.unique(lengths)
    255     if len(uniques) > 1:
--> 256         raise ValueError("Found input variables with inconsistent numbers of"
    257                          " samples: %r" % [int(l) for l in lengths])
    258 

ValueError: Found input variables with inconsistent numbers of samples: [2052, 684]

#

Hi so I got this error, I understand there are more samples and I went online to try resolve it

#

Most solutions said add .reshape(-1,1) to fix it

#

But it didn't do anything

#

😦

dapper karma Dec 16, 2020, 4:57 PM

#

storm gate its def gonna be faster then pyautogui

how would i find an image on the screen though? i think the library only allows me to look for an image inside another image not on the screen

storm gate Dec 16, 2020, 5:05 PM

#

you can screencap the screen im pretty sure

#

https://stackoverflow.com/questions/35097837/capture-video-data-from-screen-in-python/43560140

Stack Overflow

Capture video data from screen in Python

Is there a way with Python (maybe with OpenCV or PIL) to continuously grab frames of all or a portion of the screen, at least at 15 fps or more? I've seen it done in other languages, so in theory it

#

theres other ways to do it for sure