Hey all 
I built a YouTube Fear & Greed Index that uses GPT-4o-mini or gpt-4o to analyze video titles and track market sentiment trends over any date range.
Live app: YouTube Fear & Greed Index
LinkedIn: Ivan G.
Why?
I often screen YouTube channels and their sentiment to identify price–sentiment divergences or to spot potential market tops and bottoms when sentiment becomes heavily skewed in one direction — greed or fear. So the tool could be useful for those who trade/invest in TradFi/Crypto/Forex and use sentiment as part of their system.
P.S. Most of the time, sentiment is reflexive and follows price. The real alpha, if any, lies in finding these absolute extremes in sentiment. Not financial advice ofc. But more or less, this tool can be used as a substitute/confluence with the Market Magazine indicator for those who are familiar 
P.S2. All of it was built with the help of LLMs - I am a noob coder with very limited experience.
How it works (quick)
- Fetch titles for a channel between Start Month/Year → End Month/Year
- Classify each title as Bullish / Bearish / Neutral (model-only, prompt-guided)
- Compute an overall index and plot a monthly breakdown
- Export to CSV / Excel for deeper analysis
Features
Date range selector (e.g., Dec 2023 → Jan 2025)
Model switcher: gpt-4o-mini (cheap/fast) or gpt-4o (smarter)
Model-only classifier with clear rules (mixed titles → Neutral)
Big, color-coded Overall Index + monthly line chart
Labeled table of videos (with links), CSV/XLSX export
Caching on YouTube fetches to keep it snappy
How it works (quick)
- YouTube Data API v3 → playlist uploads → video IDs
- Batch query video details (duration, views), filter by min minute
- Map labels to scores (
Bullish=+1,Neutral=0,Bearish=−1) and aggregate
Tech stack
- Streamlit (UI, caching)
- OpenAI API (gpt-4o-mini / gpt-4o)
- YouTube Data API v3
- Pandas + Altair for tables/charts
