Stitch

• 

  Graph

• 

  Landing

Inspiration

We were aware of a hole that needed to be filled in the streamer economy. People are spending hours clipping themselves or others to try to capture the most viral moments from a stream and post it on social media. We wanted to streamline this task, or at least provide a helping hand along the way.

What it does

Stitch scrapes all the chat data and time stamps from the Twitch stream. Then we run this data through a sentiment analysis model, and combine the frequency of message activity at a given timestamp with the sentiment score to determine the moments in your stream that got your audience's best reaction.

This data can then be turned into links to that specific time stamp for convenience. Then all the user has to do is select download on twitch for that highlighted segment, and they have their video.

How we built it

We used the Twitch API to scrape chat data and timestamps during the stream. From there, we implemented a sentiment analysis model using VADER, which processes the chat data to identify the mood and reactions of the audience at each timestamp. The sentiment data is then combined with the frequency of messages to highlight key moments that had the highest engagement.

Challenges we ran into

We found out that Twitch API does not provide the ability to download VOD (Video On Demand) objects. Hence, we would be unable to edit the video and return a finalized video to the user, as we had hoped for.

Accomplishments that we're proud of

We persevered through numerous setbacks, adapting to the roadblocks we encountered throughout our implementation process.

What we learned

We gained valuable experience working with real-time data processing, sentiment analysis, and integrating third-party APIs. We also learned how to work within the constraints of the Twitch API and came up with creative solutions to circumvent the limitations, like creating links to timestamps instead of directly downloading videos.

What's next for Stitch

We would like to expand it to have a way to further automate video downloads and editing (adding subtitles, cropping etc.). And also including a direct share to platform of your choice within the web app.

Built With

• javascript
• node.js
• react