This is a submission for the Cloudflare AI Challenge.

What I Built

I built an simple application to look up any popular place and quickly summarize and understand the sentiment of the place using reviews data

Demo

https://placeinsights.pages.dev/

My Code

https://github.com/problemsolver03/place-insights-cloudflare

Journey

The idea was to utilize the Google Maps review data and use it as the input data to summarize the reviews and also to provide users with the sentiment of the consolidated reviews.
This quick summarization and sentiment classification will help users quickly get an overview of the place they want to visit.

Once the idea was clear I procured a Google Maps API key to get the reviews of the place.
then I created two workers on Cloud Flare one for review summarization and the other for analyzing the sentiment of the reviews, with these workers setup. I was able to pass the review data to the workers and get the output of the models through API response, which I used in my react frontend to visualize it and later hosted the react application on Cloudflare pages.

Technologies/Libraries used
React, Serverless functions, TailwindCSS

Multiple Models and/or Triple Task Types

bart-large-cnn
distilbert-sst-2-int8