PERSONAL PROJECT · EXPERIMENT
Trend Radar
Turning a stream of public signals into a more deliberate creative process.
The question
Can a small research pipeline help me find interesting topics, explain why they might matter, and run better creative experiments?
What I built
A personal pipeline that combines TikTok trend observations, Google Trends, and YouTube signals. It collects snapshots, compares signals, and uses AI-assisted analysis to organize candidate topics for me to review.
- InputPublic signalsTikTok trend observations, Google Trends, YouTube
- StepSnapshot collectionKept stable so runs can be compared
- AICompare & analyzeAI-assisted analysis organizes candidate topics
- PersonChoose & approveI pick the concept and approve each clip
- OutputPublish ShortsCollection and publishing stay separate
- StepCompare with responseCreative choices checked against actual audience response
Collection and publishing are separate. I choose each concept and approve each clip; the analysis score is one input to that decision.
What the experiment showed
At the October 17, 2026 checkpoint, three YouTube Shorts had each passed 25,000 views, and the channel had 51 subscribers. It’s a small sample, and I can’t credit the views to the pipeline alone.
Results varied a lot across platforms. The opening visual and replay behavior look like the most promising things to test next.
What I’m learning
Keep the collection process stable, separate hypotheses from evidence, and compare creative choices against actual response. A compelling signal can still produce a weak result.
KEEP IN TOUCH
Questions, ideas, or a shared interest?
I’m always glad to compare notes on practical software and AI automation.