Two years ago, I decided to start working on a Open Data guide for Python here Open Data SE

The idea was to collect with the help of the rest of the users all the possible file formats that an open data researcher will may need to work and add a Python library with an example.

It proved a success story for the Open Data since it is one of the most popular questions. As a result, I want to do a similar concept here in Data Science. But, I want your feedback if this will be accepted since I am not in this SE from the starting date.

I will create a question like the one in Open Data asking about different machine learning algorithms and in which cases it can be used. Then, I will start an answer which I update all the time with comments' feedback and stuff that I find.

I will use as a starter this post from Kaggle and populated with links from Wikipedia. Then, the idea is to find tutorials from each category and add them too. As a result, we will end up with a great resource for every Data Scientist. Newcome or old one.

What do you think?

EDIT: It looks that the Kaggle page has been removed. Since it was a nice one, I will add it again from web archive

  • $\begingroup$ The Kaggle link gives me a 404. $\endgroup$
    – Erwan
    Dec 17, 2020 at 2:30
  • $\begingroup$ Added it again from web archive. The page has been removed. $\endgroup$
    – Tasos
    Dec 17, 2020 at 9:26

2 Answers 2


This sounds like a community wiki, if anything. I think it's possibly OK, as long as it's not just copying links from other similar wikis. Let's figure out why this collection of resources would be uniquely useful on StackExchange.

  • $\begingroup$ Then, I could spend a couple of days on this to create a draft based on the Kaggle post, then post it here and as long as I have your green light, I post it on the regular page as well. Is that ok? I didn't want to spend time on this if it would be a definitely no from the beginning. $\endgroup$
    – Tasos
    Dec 1, 2015 at 10:37

I think it's a good idea, imho the site could use a few "signpost" general questions like this.

I'm a bit concerned that this could turn into a super long list of very specific methods, maybe it should be restricted to the most common use cases? Or maybe divided into several questions such as:

  • Main methods for text classification
  • Main methods for image classification
  • Main methods for clustering
  • Main types of neural networks
  • Main methods for learning from sequences
  • Main models for time series
  • ...

It's just an idea to organize things a bit, not sure if it's a good one ;)


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .