#Need help in a DataSet Analysis Problem

16 messages · Page 1 of 1 (latest)

zealous glade
#

Hi, I am a New Learner. Can you help me in this problem. I don't know how to get started. I would appreciate any help from your side.

Screenshot_2022-12-21_at_1.50.24_PM.png
loud gulch
#

Do you have more context? Or is that it?

zealous glade
loud gulch
#

What does the data look like?

#

You really should provide all relevant information when asking a question

potent salmonBOT
#

@zealous glade

nikkhil.pareek Uploaded Some Code

this is the csv file

Uploaded these files to a Gist

population.csv

View The Gist

For safety reasons we do not allow file attachments.
loud gulch
#

@lethal shard I'm about to sleep but maybe you'll be interested in helping with this question? If not, I can try tmr

lethal shard
#

Hey there.

I'd say the problem description is a bit vague, but the way I understand it you're supposed to estimate the population in a future interval and calculate the error with respect to the 2030 population.

The first thing I always recommend in these kinds of problems is to plot the data and visualize it. It will give you quite a few insights on how to approach the problem. Definitely do that before anything else.

I might be wrong here, but it seems the kind of model you choose to estimate the population isn't all that important in this exercise as its objective seems to get you familiar with model training and evaluation. So I'd assume you're free to use whatever you're comfortable with.

If you need a pointer, decade world population projection can be roughly estimated using linear regression. It's simple to implement in whichever programming language you are comfortable with. But if this feels like a daunting task to you, try using libraries that abstract the math-heavy stuff from you. I do recommend eventually implementing it from scratch, however.

@loud gulch can correct me if I'm pushing you in the wrong direction.

loud gulch
#

Yeah tbh, I agree with Miihr, it's a little vague and confusingly written

#

Do you know what statistical method they're referring to? Is there a specific way they're expecting you to do this or nah?

#

@zealous glade

zealous glade
#

I asked my Professor he told me to use linear regression

loud gulch
#

That would make sense

#

I would start with plotting the data you have at the moment as Miihr suggested

#

From that, it may be easier to see what direction you'll want to go in regards to modelling