VizChitra 2025

{Describe your talk/session in 2-3 paragraphs}
The number of hours you spend on social media on your phone is often the only metric that is used to paint a scary picture of social media addiction. The accompanying visualisation on your phone’s OS with this number is usually a time-based graph of the amount of time you spend on a particular social media app over a week. While this number/graph can cause guilt, we know that we go for social media for more than just wasting time. We want to stay up to date with advancements in our chosen field of work, to stay connected with friends/peers and of course to also be entertained.
We can do better by automating multimodal analysis of our social media feed to analyse and visualise our social media browsing patterns and make better informed decisions on content consumption.

As part of this project, we have created workflows to scrape the data on our Instagram Reel’s “For You” Page and used Machine Learning to process the videos and group them by thematic and visual similarity.
We’ve then tried to use various techniques like t-sne to create 2D visualizations of this video data to help understand our brainrot better. Here is a link to a video showcasing our visualisation.

With this talk we want to share our early findings and get feedback from a community of practitioners on how to improve this too.

{Mention 1-2 takeaways from your session}
The number of hours you spend on social media does not have to be a scary figure if you had more insight into the breakdown of various type of content you spend time on and had ways to control your consumption.

Multimodal data that changes over time requires non-standard visualization techniques which might not be easily understood by a general audience. Work needs to be done to make it easy to create such interactive visualization and to make it easy to understand for the viewers.

{Who is the audience for your talk/session?}
Nerds of various feathers - Data Visualization Social Scientists, Computer Science, Machine Learning and fellow Brainrot Victims.

{Add your bio, including work place name and your job role}
I work as a Software Developer at Tattle Civic Tech.
We at Tattle build open-source tools and datasets to respond to inaccurate and harmful content online in India.