College athletic teams and related organizations often look fail to look into the social media arena when looking at potential hires/recruits. Online behavior of individuals is an indicator of a person's character, and this could be a way to analyze the nature of your recruits to see if they'll be a good fit!
What it does
Insight analyzes social media and online behavior, and makes all the results accessible in a simple, user-friendly way! It performs sentiment analysis, entity recognition, tone analysis, and more on users' twitter profiles to make it a valuable resource for the character evaluation process in recruiting.
How we built it
We use IBM Watson's Natural Language Understanding toolkit to analyze tweets from an individual. We clean the data, extract useful information, and display it on a dashboard. This dashboard is on a web-app and makes use of React-Redux, and D3.js.
Challenges we ran into
Choosing qualitative metrics whose data could be obtained by NLP models. Also, for a given name and hometown, there are multiple possible social media profiles corresponding to different people. In such a case, it was imperative that the user provide the social media website link to avoid erroneous reporting.
Accomplishments that we're proud of
Applying NLP methods to extract meaningful information, and also building a web-app with an elegant dashboard in an relatively short amount of time!
What we learned
We learned a lot about the applications and potential of Natural Language Processing in helping us solve actual problems. We also improved our technical understanding of system design, and learnt some cool new technology along with it. In addition, it was interesting to see how a candidate's personality and fit is assessed before recruiting decisions!
What's next for insight
The next step would be adding more data input sources, which would need permissions and authorizations amongst other things.