Feedback Analyzer

analytics
dashboard
tweets
more tweets

Inspiration

We were inspired by the RBC challenge presented to us during the sponsor speed dating event. It interested us because it forced us to go out of our comfort zone to learn new things and work with technologies we had only heard of before.

What it does

It takes tweets that mention RBC's "@AskRBC" twitter handle and categorizes them as technical or non-technical tweets. Then, by using natural language processing we can check to see if the tweets are "positive" or "negative" with up to 99% accuracy! We then compare the tweets with real time statistics from RBC's website to check if the website is up or has any delays.

How we built it

We first built a front-end for our web app using Bootstrap, HTML/CSS and Javascript. We then made a back-end for the app using Flask, with various cloud technologies such as a SQL database hosted on Amazon Web Services. Finally we wrote scheduled scripts to automate the process of getting tweets and categorizing them, applying natural language processing and to check RBC's website status.

Challenges we ran into

Since we are first years, it was quite challenging for us to work with these technologies for the first time, and took a lot of effort to learn and apply the skills in under 24 hours! The Twitter API is very frustrating to work with, we ran into rate limiting issues and in some cases, we could not do as much as we wanted to since the API is mostly locked behind a paywall.

Accomplishments that we're proud of

We are proud to have completed our first Hackathon! And hope to see many more. Getting the database setup on Amazon Webservices, and making sure it was working properly and able to communicate with our application was a big achievement. Also, getting all the tools and technologies we used to work together was quite the challenge. All in all, we are excited to present our achievements to the judges.

What we learned

We learned how to use natural language processing with a web app, to improve the ability of a business (in this case RBC) to help it's customers. We picked up a lot of new skills on the go tonight, and taught ourselves various things; from simple programming to advanced machine learning and AI topics. Being first years, we learned a lot.

What's next for Feedback Analyzer

We plan to improve upon Feedback Analyzer by adding more social media platforms, more statistics and analytics and hopefully turning it into a startup!

Built With

Submitted to

cuHacking 2020

Created by

I worked on the database, getting the tweets from the API, using a fetcher that periodically got the latest tweets and stored them categorically in to the database. I also worked on the front-end and designed the website with Bootstrap. In the back-end I worked with Musa and integrated Shane's machine learning NLP model into the Flask app.

Hamza Sohail
I worked mainly on integrating the front-end and the back-end. Along with creating the initial Flask web app and configuring HTML/CSS files, I retrieved tweet information from our SQL database in order to properly analyze and neatly display the tweets and their respective sentiment analyses.

Musa Ali
I added NLP and Machine Learning to the project. This helped us with classifying the different types of tweets whether it was a 'positive' or 'negative' tweet, and whether it was a 'technical' or 'non-technical' tweet from a list of keywords.

Shane Fernandes