Seeing how the Microsoft sentiment API was able to take a message and output the Sentiment value of that message, we got the idea of using that same API on facebook messenger. We have all had moments while talking on facebook where we just wish we knew what mood the other person was in. For example, when flirting with a girl over messenger, we want to know what mood the girl is in and that's when this extension would come in handy.
What it does
Messenger Sentiment takes in the latest facebook messages in a chat and utilizes Microsoft Azure Tech Analysis API to assign a sentiment score of 1-100 for each message, with 1 being most negative and 100 being most positive. This extension then assigns each messeage one of 5 emojis to represent the tone of the message.
How we built it
We built a google extension that injects the current facebook messages to our php client hosted on heroku which does a post request to the microsoft azure api.
Challenges we ran into
One of the challenges was getting the live data from facebook messenger into a text array to be sent to the API. Another challenge we had was client privacy issues regarding sending a post request from messenger.com. A third challenge was setting up a https php application that redirects the post request to azure's api. We also had an issue with scaling the application due to the api's limitation of making only 100 calls per minute. Our last existing issue is scaling this application due to the cost of using Microsoft's API.
Accomplishments that we're proud of
Our biggest accomplishment was getting our extension to work and bypassing some online security measures.
What we learned
What's next for Messenger Sentiment
We could push it to the google extension store to be used by the first 5000 calls every month.