Example: Building a Sentiment Analysis Dictionary for a Non-English Language

Most questions here are about retrieval or processing of data. Should data compiling be considered on-topic?

Personally, I would like to see the site focusing on use of open data, because data collection is a vast and rather different area.

  • 1
    As long as the question is about collecting open data, what's the harm? – Patrick Hoefler May 9 '13 at 12:30
  • Do you include data extraction techniques in data collection? or don't you consider the techniques? – Vince May 9 '13 at 13:27
  • 1
    @Vince What I meant by data collection is collection of raw data (conducting surveys, developing Internet crawlers, for example). Extracting information from an already existing data set is not what I intended to include. – Andreas Blaesus May 9 '13 at 13:42
  • See also the related discussion at Will we allow questions that deal with computational linguistics corpora? – Patrick Hoefler May 9 '13 at 16:05
  • @AndyShuXin Note that my answer involves extracting/synthesizing new data sets using existing data sets. – Kyle Strand May 10 '13 at 1:37

I would probably vote to close something purely about data collection. I think information about sources and about how to use sources is on topic.

I'm also kind of receptive to posts about best practices about publishing data, but I still think collection is a step too far out.


I don't think it should be closed. I think we should look at the bigger picture. Open Data is quite focused on the "open" part, so on the release and collection of data. There was a short discussion talking a bit about it.

There is a growing data.SE project, SO and some other projects probably have many questions on the topic of analysing and manipulating data (even though I don't know how wide the scope of your "use" is).

But I agree that there is a large part of the questions that will cover data collection, and it should not pollute the rest of the site. That's why we should tag them properly (and maybe have a template). There is another discussion about it.

  • 1
    Ah, pardon me for my vague wording. I meant to point out that the example question is about construction of a data set that doesn't previously exist. While other data-collection questions are about locating data sets that does (presumably) exist. – Andreas Blaesus May 9 '13 at 12:18
  • @AndyShuXin Ok I see. I guess I'll rephrase my entire answer. – Vince May 9 '13 at 13:25

You must log in to answer this question.

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