We took the problematic from Bancomer. We are showing in a heat map the possible consumers for Bancomer. We are getting all this data via social networks (Twitter) and machine learning.
What it does
We have 3 services. The first one is an application in Java that can run like a server and has an endpoint where you can call it to classify the tweets. The second one is a web interface to show in a heat map all the data that we get in the third service that it is the one that connects the interface and the Perceptron, and also stores the data in a DB.
How we built it
For the Perceptron we used the Weka API, Spring and Java. For the Web Interface we used Vue.js. For the Middleware we used some packages from npm.
Challenges we ran into
The time was the main issue but we managed to accomplish all the goals that we planned.
Accomplishments that we're proud of
Web service in Java, the complete trained Perceptron with Weka, the scripts for running in a faster way Weka, the interface with Vue.js and finally the integration.
What we learned
We managed to learn new technologies like Spring, Weka and Vue.js
What's next for FBs-HeatMap
We want to integrate more social networks in the Middleware.