The below cluster graph was part of an analysis of AI influencers and their conversations about Generative AI and Large Language Models. Here was the process I took:
- I started the analysis by uploading a CSV file of 18,000 social media mentions of AI influencers about generative AI into Claude.
- I used Claude’s Sonnet 4 to analyze the data and create the artifact.
- I downloaded the code of the artifact and saved it as an index.html file.
- I then used Cursor to upload the file directly to a new repository in GitHub and then published it to Vercel.
Using the AI creator data, use topic modeling to identify 5-7 distinct narratives and cluster them graph. I want to understand how each of the narratives are connected from semantically relevant keywords and themes.
What’s interesting about this process is that it was much easier to work with the code in Cursor because there were some small tweaks I wanted to make. First, the cluster graph was not centered on the page. Also, the title and the navigation were not responsive, and they were also justified to the right. After some back-and-forth with Cursor’s AI agent, I was able to fix this and get it uploaded to Vercel within 45 minutes.
