Keyword enrichment analysis of published research articles

In the ressource and analysis you will find below are summarized an idea I had about trying to use the keywords of published research articles indexed in Pubmed and try to see whether trends would arise or not. In other words, can published research article keywords indicate modifications of research interests globally ?

This a first step, aiming at :

• Get as much data as possible from pubmed on a rather sluggish laptop running a flavored Archlinux
• Integrate this data into a clean database (MongoDB)
• Start to explore the data “by hand” and try to get an intuition whether this whole idea make sense or not at all
• Does journals with a given impact factor modulate the presence of certain keywords ?

All ressource code can be found on Github.

The next step will be to crunch some real numbers beyond the intuition, but by then the following Jupyter Notebook gives my results so far :