CLS- Graph AI

Inspiration

My interests is doing Data Science for Social Good. Covid 2nd wave has killed many lives due to the unavailability of basic needs which inspired us to build a system that could save lives.

What it does

. Tweets are categorized into : Request - people requesting for Help on Twitter Offer - people Offering Help on Twitter Medical Help - Tweets related to medical help

If a user is interested to offer help,They can filter by requests and further filter by location and help.

Also,There are different kind of new medical help changing over time . As pandemic progressed, there was new requirements like Injections,blood plasma etc. So, we developed a Semantic Search so that users can get related tweets based on User's Query.

How we built it

Collecting tweets using tweepy python. Run Multioutput Classifier to classify tweets into Request,Offer,Medical Help,Medical Products,Aid Related Etc. Use Semantic Search to get related tweets based on User Query.

Challenges we ran into

We used d3.js to get Network Graph.We tried to get the network graph interactive , node zoom based on user clicks and it was very challenging to get that.

Accomplishments that we're proud of

Classification Algorithm with 93% Accuracy Model classifies exactly the people trying to request as "Request" and other categories as well. Getting the tweet object within the table which is very easy for user to click and respond directly to tweet. Learning GraphQL and Implementing the GraphQl API's.

What we learned

We Learnt GraphQL and it was really a nice experience.We have gained knowledge on GraphQL and the power of using GraphQL and will be exploring more and using in all our future projects.