1
$\begingroup$

I have 2 months of air pollution measured data for a location of interest. I am highly interested to play with machine learning algorithms to predict pollution at the same location (let say, for the next two months).

Since air pollution is a geographic problem, does it make sense to apply and predict pollution at a single location using machine learning? Could someone shed a light on prons and cons of this?

$\endgroup$
3
$\begingroup$

Not enough karma to simply comment, so I have to log this as an answer.

Does it make sense? Sure! But keep in mind that if you train your model on a single location, it can only really be trusted to be used on future data from that same location. Could your model generalize to other locations perhaps? Maybe, but your model didn't see anything so you can't be sure. Same thing applies to having only 2 months worth of data. You can't be sure if this model will be accurate when it's scoring data that comes from a month it hasn't seen before.

If you really want to work on this type of problem and just don't have the data for it - this link has a similar dataset that might interest you.

$\endgroup$
1
  • $\begingroup$ Thanks for the answer and link, I meant to play with the data and predict pollution for the same location. I agree it wouldn't make sense to use the same model for other locations. $\endgroup$
    – khajlk
    Jun 25 '20 at 10:59
2
$\begingroup$

Imho with only 2 months of data one big problem you will have is seasonality: pollution depends a lot on temperature and wind conditions, so in order to predict pollution level accurately you would need at the very least one year of data, preferably several past years. I don't think you can predict properly pollution level two months in advance with only two months of training data.

Maybe you could try to make the problem broader: use a dataset including multiple cities with several years of data, then you can try to forecast pollution level in a particular location using features representing local conditions such as density, temperature, etc.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.