Twitter Replay

Twitter Replay's home screen
An example of a streamer's relevant Twitter trends

Inspiration

We are both avid fans of various Twitch streamers and greatly enjoy the opportunity to interface with the online community both through their streams and through their Twitter fanbases. However, missing out on real-time streams and watching video-on-demand copies (VODs) often leaves out this social aspect of Twitter interactions. We offer a way to watch Twitch VODs without losing the ability to experience Twitter trends in real time, allowing users to feel more connected to the Twitter community associated with their favorite streamers.

What it does

This website offers a way to "replay" Twitter trends that relate to a given streamer. Users may input a link to a VOD and will be shown a list of Twitter trends that occurred during the livestream that included Tweets that often made reference to the streamer.

How we built it

We created a front-end Python interface that depends on a SQLite database that stays updated with the latest Twitter trending keywords and finds non-filler words (removing stop words like "and", "but", "or", "her", "his", etc.) that often occur next to these trending keywords. When a user requests a VOD on the front end, the web player is able to display the VOD and searches the words commonly associated with these trends for mentions of a streamer's name. If any are found, the trend is highlighted next to the VOD player and the user can both see when the term began trending and may also click on the term to be taken to that timestamp in the VOD.

What we learned

We each learned a lot of about SQLite, Twitter scraping, and the connection between front end and back end development. This sort of collaboration was new to both of us and allowed us to sharpen our respective skills through a project that drew on our personal interests.

What's next for Twitter Replay

Next up is streamlining the interface and potentially adding speech-to-text in order to provide video markers for quoted clips (for example, if "gummy bear" is trending alongside the streamer name "Shroud," we'll be able to find the point in the VOD where that phrase is first said). We hope to expand this website into a full-time project, so stay tuned!

Built With

Submitted to

HackNC 2021

Created by

I was responsible for putting together the Twitter interface through the Python library tweepy and sending all relevant data to the SQLite database. The former was familiar to me but the latter was a challenge, as I've never worked with SQL before.

Nolan Welch
I created the front end for the project, using the Twitch API to get relevant data for the selected VOD and SQLite to access the data gathered by the Twitter bot. I also used Flask to deploy the website to display the data. I was very familiar with all of the tools I used, however putting it all together into one cohesive project did prove difficult.

David Foss

Updates

Nolan Welch started this project — Nov 07, 2021 11:37 AM EST

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.