#🔒 ARIMA and times series weather forecast

32 messages · Page 1 of 1 (latest)

visual ironBOT
#

@proud scroll

Python help channel opened

Remember to:

  • Ask your Python question, not if you can ask or if there's an expert who can help.
  • Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
  • Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

Closes after a period of inactivity, or when you send !close.
proud scroll
#

  1. These 3 CSV files have a total of 9 years of historical data about Finland's weather on 4 factors (air temperature, cloud cover, precipitation, and wind speed).
    I tried combining the 3 files and it messed up the columns and everything so it couldn't be read. The reason why they are in 3 files is because the official weather site allowed me to download up 3 years spans.

  2. The 3 files need to be trained and tested with ARIMA

  3. Using Time series analysis with the 9 years of historical data I have to get charts showing forecasts for the next 10 years. Come to a conclusion, is climate change increasing or decreasing based on the analysis of the 4 factors throughout the 4 seasons (Summer, Autumn, Winter, and Spring)?

I have been on this for days and the code just doesn't work. I am not sure if the CSV files are not understandable. The forecasted data is not realistic and it is showing a straight line.
I am using google collab.

peak ether
#

First, deal with one problem at a times if your data is bad, no point moving on to next step.

#

Plot your data over time, and examine it to see if it makes sense

#

Once you're there, then you should look at using pmdarima (and its auto_arima)

#

Then, you need a plan for handling seasons. You could use seasonality in pmdarima.... or you could calculate a seasonal average or distribution and forecast that independently

proud scroll
#

I need a step by step guidance.
By ''data is bad'' you mean the csv?

peak ether
#

The first thing you do with any new dataset is plot it: make sure it makes sense.

proud scroll
#

These are official data from an official weather site. The CSV looks funny, but it is the only data I can use for this

peak ether
#

Great, but does the data make sense?

proud scroll
#

the information in it yes

peak ether
#

'The csv looks funny' is not very informative.

#

Funny how?

proud scroll
#

Well it is a subjective matter, if you look at it, it isn't the typical American type column setup

#

in my opinion it looks funny

#

others might disagree

peak ether
#

Still not descriptive.

proud scroll
#

The thing is, I don't know how to explain it

#

It is something that you see and you realize if it looks proper

#

But because it is from the only official site, It has to be it

peak ether
#

I just want you to appreciate that nothing you're saying is technical nor descriptive. There's nothing to respond to.

proud scroll
#

the colums look like they are stuck together or all written inside the same colum

#

but the code still understands them as seperate

peak ether
#

Instead of stats model, I'd also suggest using pmdarima https://pypi.org/project/pmdarima/. It contains a very helpful/important auto_arima feature that discovers p,d,q's for the arima model

peak ether
#

That's why you plot and inspect the data first

proud scroll
#

I am currently testing pmdarima

visual ironBOT
#
Python help channel closed

This help channel has been closed and it's no longer possible to send messages here. If your question wasn't answered, feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.