Expand Scope of Acceptable Data Science Questions?

Question

This is my first time posting on meta so please excuse my question/comment if it needs to be changed in any way.

I am a regular user of SE Data Science but recently have discovered that a lot of questions on the "Orange" (Data Mining) tag are being closed for being off topic. When I read over some of these questions, to me at least, some seem like they would fit well on the platform.

Data Science is a very broad topic which I think in some cases can include topics like finding data, data mining, and visualization. I know other sites exist (like Open Data) but should topics like data mining and visualization be included in the list of acceptable questions listed on the site?

Would it help to add a page that provides a list of acceptable topics to ask on the site?

I found some similarity with a question that was recently asked here: What is on-topic here?

There is a lot to unpack here but it was just something that I noticed.

Answer 1

As I understand, Orange questions are within the scope of this site as long as there is a good explanation and it's related in some way to data science. The Orange tag has had 9 questions asked this month and 72 this year so people see value in said questions, most of which have answers and only one was closed this month (probably two in the future). The closed question was "why is there no volcano plot in the new version of Orange" which is really not relevant to the site / low effort question.

About data mining (16 asked this month, 252 this year), the way I see it is that the tag and questions are relevant to the data science site because oftentimes we are facing doubts about the results and one of the main questions is: "Maybe I messed up something when I was gathering the data", or even problems related to the format questions that people see as relevant on this site given the answers given on said questions. General API or scrapping questions shouldn't belong to data science, something like general BeautifulSoup questions or general Scrapy questions.

Open data is mostly for data source requests so data mining and code don't really belong there. I guess questions about the correct use of APIs would be relevant there as well as on Stack Overflow.

Programming questions should follow the same approach in my opinion. Numpy, Pandas, Sklearn, Keras, etc. are the most commonly used libraries in data science and there will be naturally questions related to them. As long as the question is related to a problem in: processing data, training-testing models, visualization of data and models, storing models and data, ensembles, etc. they should be seen as on-topic. This can be exemplified by: keras Sequential CNN for image data reshaping data issues here a question about how code is directly related to a problem in CNNs and reshaping of data (pre-processing). R is mostly used for data science, ML and statistics and not for much more so it's hard to find an unrelated question in R in my opinion.

What I think it is an interesting proposal for expansion in my opinion are questions about doing data science in the cloud, like, in AWS instances or VPSs. My guess is that more and more people will find problems specific to the use of cloud services and still will be related to data science.

Answer 2

What's an example of a question you would or would not consider on topic here that has been closed or left open? Usually programming questions are moved to SO even if about things like numpy or R. I imagine some data mining questions fall into that category