We want to use ML and DL for the social good. We want to use NLP, ML to reduce the burden on the underlying Health Inspection Machinery by the government, to identify key restaurants with health risks.
What it does
A small city has approx. ~10000 restaurants (registered) and countless many undregistered. With these many restaurants comes a responsibility of assessing the risk associated with health in these places.
Worst part is that the government just doesn't have the machinery, to visit every restaurant. A normal city, if at all, has ~100 assesssors. It's impossible for them to cover it all!.
WHAT if, we could help them solve this problem by the power of Data!
How we built it
We develop a risk prediction system using the customer ratings to generate the rating of the restaurant on a 1-5 scale with 5 being the safe ones and 1 being the worst ones!
We can identify the key restaurants that need a visit by the assessors to identify the health risks. That way, we can ensure that the minimal resources are used where they need.
Challenges we ran into
It's complex! integrating NLP with sentiment analysis, and with ML, and that too in a matter of hours! (Our last project failed!)
Accomplishments that we're proud of
Generating the ratings and the underlying business specific data to score all the restaurants!
What we learned
NLP, sentiment analysis, python, GPU training!
What's next for Public Policy: Health risk assessment using ML and NLP
Expand the model to using live twitter feed to generate a alert system for the assessor to get suggestions in real time.