# CrossValidated vs DataScience: what is different? [duplicate]

Probably it will be off-topic, but anyway I really want to know the difference between this forum and Data Science forum? The both communities are the subparts of StackExchange and at the moment I don't see the reason to split the Data science questions in the two parts (consequently to provide two different communities)...

• My guess it's probably the name. CrossValidated is quite an obscure name. A casual browsers might not see the connection to Statistics or to DataScience. Its a shame really, because good Data Science must absolutely be concerned with the theory and application of various probability models & statistics. – Assad Ebrahim Dec 3 '14 at 8:31
• @Assad, CrossValidated is obscure but Data Analysis & Statistics is not. It is pity to some extent that new fashionable disciplines such as Data Science and, earlier, Data Mining, try to promote themselves at the expense of Data Analysis by partly duplicating it. – ttnphns Dec 3 '14 at 9:09
• I suppose this will soon be obsolete because DataScience.SE seems bound to fail (unless something unforeseen happens). See area51.stackexchange.com/proposals/55053/data-science – Momo Dec 4 '14 at 21:13
• Cross Validated vs Data Science: i.redd.it/5193db0avbey.jpg – Imran Dec 18 '17 at 14:53

In my opinion cross-validated concerns the theory and methodology of analysing data, e.g. statistics and machine learning.

Data Science, on the other hand is more hands-on and deals with implementation related topics such as data handling, software infrastructures, databases and toolchains. This is particularly true at the start of Data Science.

That said, there is an increasing gray zone between the two. Data Science is receiving more and more questions that, in my opinion, should end up here instead.

• My problem is, I'm relative active on the DS forum, anyway I would like to discuss some question, which were published here. For some users (like me) it is difficult to jump between two simillar forums with the small numbers of topics. Probably it would be good solution to combine both in the one? – Guforu Dec 3 '14 at 8:36
• @Guforu, I don't think combining will be a good idea. Rather, the two should put it clear in what respect they demarcate. CV is data analysis (broad, including also learning/prediction and mining); and I even think CV ought to be renamed accordingly. CV doesn't accept data management questions which, I suppose, is the core topic of DS. – ttnphns Dec 3 '14 at 9:14
• @ttnphns not sure ... I think a problem of CV is that it has a higher barrier of entry compared to data science for people without a theoretical background. We see a lot of theory questions popping up at DS and usually plenty of people can answer said questions there. I think trying to change that dynamic won't work. I'm not sure that keeping them separate is the best choice in the long run. Keep in mind that CV is not that active on its own. Maybe one large 'data' SE would be better than two smaller, intertwined fora. Stackoverflow shows that a wide variety of topics can work fine. – Marc Claesen Dec 3 '14 at 10:01
• I have to wonder about SO. In some ways it looks dysfunctional to me with $\sim 10$K votes to close (not so long ago it was more like $100$K, until the number was cut arbitrarily). It works not so badly when it behaves as almost closed sub-communities. Presumably surveys show that many people prefer all the questions of some interest to them together in one place (and all the others elsewhere). Setting up DS was based on a deliberate decision that such questions don't fit in here, it seems. – Nick Cox Dec 3 '14 at 20:09