I was just inspired by seeing other sorts of natural language processing demonstrations online. I thought it would be fun to build one.

What it does

This presents a simple dashboard for administrators and social media managers to quickly see possible issues with their production environment. It leverages Google Cloud's extremely powerful Natural Language API to track customer attitude towards RBC on social media.As well, it allows for the collection of data so that overtime we can see if we can draw correlations between the datasets of production issues and poor sentiment performance on social media.

How I built it

I used a Python w/ Flask backend and a React frontend. I used Google Clouds Natural Language Processing API to take sentiment data of tweets that I scraped with Tweepy for python.

Challenges I ran into

There were lots of challenges that I ran into. I spent a large portion of my time on making the backend pipeline. I was having trouble deciding on which technologies to use.

Accomplishments that I'm proud of

I think that the UI looks quite nice. I generally struggle with design but I'm pretty proud of the overall look of the dashboard.

What I learned

I learned a lot about the inner workings of Googles NLP engine, quite a bit about using APIs such as Twitters, and a lot more about Flask as a backend

What's next for SDC (RBC) Issue Tracker

Quite a lot of bug fixing. As well, I think it would be interesting to expand the reach of the product into other companies. I think the possibilities of using NLP on social media to track production issues are quite large and if I were to expand the query into companies outside of RBC quite a lot of interesting data could be gleaned. As well, working with other languages would be interesting and beneficial. I think French would be something easy to implement.

Built With

Share this project: