As a potential Ph.D. candidate in data analytics and as an AI enthusiast and architect. I wondered why is it not the case that I can ask an AI that question "Which research topic would you recommend me to dig in"?! A bit vague though, so I decided to be more specific in order to simplify the problem for the AI. Which research topic is getting most grant funds?!
I think an AI should be at least able to answer this question if you give it a data source that it parse/read from which would have all the grants stored. Hence the question what about an AI, that would take the original question "Can you recommend me a research topic". and modify it to "Would you like to ask me, which research topics are getting most funding and therefore most interest?!"
I decided that maybe instead of waiting for or searching for, I should start building that AI. I started with my university George Mason.
Ph.D. candidates survive on the money funded by research grants, if you are smart, you can get into a Ph.D. program, but getting into the wrong research topic with an advisor who does not win grants is a huge issue. It just means that you will not have enough funding to do a proper thesis, which is mostly the core of any Ph.D. journey. The remedy is actually quite simple, know which research topics are getting most grants and find the topic nearest to your background to work on. Grants Research are published on any university website. So where is the problem! here it is. They are not presented in any informative sense. Of course, an AI that would parse it can tell you which ones are most relevant and interesting, but if you expect a human mind to make sense of this, at least it should be visually informative.George Mason is my university of interest and that is why I focused my experiments on it but the experiment here is applicable on nearly any university.
I scrapped the data and revisualized on a very quick experiment: https://public.tableau.com/profile/emad.mohamed#!/vizhome/Grants_6/Dashboard1
Please hover over the lines in the Tableau Dashboard to see the grant details.
Easily now I can see the outliers presenting the topics winning most industry attention. A lot more can be done and much more text wrangling can be used to answer further questions or ask better ones. This is a just for a proof of concept. That the most valuable information can be out there but makes use of it because of the way it's displayed. Just knowing which topics are hot filters out a lot of bad/poorly funded topics, which alone can prevent putting effort and time with little return on investment.