It looks as though Python is starting to dominate as the data science language of choice.

I started small research on this after a friend pointed me to this blog post about trends in data science languages.

The blog post is from March 2018; here is some more recent data, as of February 2020.

This is the result of my search of Google Trends for various programming languages combined with “data science”. The graph units are as a percentage of the highest value from any of the included searches.

Here are some matching Stack Overflow trends. I added a search term for the Pandas data science library for Python. The Stack Overflow (SO) trends are for single search terms, not in combination with “data science”. For example, the Python trend line includes questions for all uses of Python, including web development and games. I didn’t know how to do the combined search, but if you do, please let me know.

The plots show that R and Python were growing at a similar proportional rate up until around the end of 2016. After that, R’s “data science” Google search flattens off. SO has a slightly different pattern; the question proportion for R continues to grow until 2018 and then flattens. Around the same time, grown in SO Python questions increases.

By the beginning of 2020, there are more than twice as many Google searches for Python and “data science” as there are for R and “data science”. On SO, the proportion of questions about Pandas – a single Python library for data science — gets close to the proportion for all R language questions.

Google trends shows that languages such as Java, Matlab and SPSS are distant also-rans for searches with “data science”.

For some reason, Matlab questions on SO peak in 2015, and start to drop thereafter, so there are many more questions about Pandas – the library – than there are for Matlab – the language.

Both Python and Java are very widely used languages in other domains, but Java questions on SO have been steadily decreasing since around 2015, while Python questions have been steadily growing, overtaking Java questions around the beginning of 2017.

Share on: TwitterFacebookEmail



Published

Category

teaching

Atom feed