Public data sources for high-attention case studies¶
These sources are useful when you want a demo or benchmark case that feels timely, public-facing, and easy to explain.
GitHub stargazer history¶
Series type: daily attention / adoption
Loader: load_github_star_history
What you get
- daily new stars
- daily cumulative stars
- repo metadata in frame attrs
Good for
- repo virality
- open-source launch studies
- attention spillover around announcements
Notes
- Uses the public GitHub stargazers endpoint with timestamp-aware media type.
- Great for high-attention open-source case studies such as OpenClaw-like repo surges.
GH Archive watch-event history¶
Series type: hourly or daily GitHub attention
Loader: GH Archive / external preprocessing
What you get
- public GitHub event stream
- watch events at scale
- cross-repo comparison candidates
Good for
- ecosystem-level star studies
- fake-star filtering research
- broad repo panels
Notes
- Useful when the native GitHub API is too rate-limited or when you want many repos at once.
CoinGecko market chart¶
Series type: crypto price / market cap / volume
Loader: load_coingecko_market_chart
What you get
- historical prices
- market caps
- 24h volume
Good for
- crypto event studies
- listing news
- policy or ETF windows
- attention/price co-movement
Notes
- Daily data is a natural fit for event windows and quasi-experimental impact analysis.
FRED commodity series¶
Series type: macro and commodity daily levels
Loader: load_fred_series
What you get
- date/value pairs
- named public series such as WTI and Brent
Good for
- oil shock studies
- macro controls
- commodity event windows
Notes
- WTI and Brent are especially useful because they naturally form treated/control pairs or spread-style controls.
LBMA gold via DBnomics-compatible CSV¶
Series type: daily gold price levels
Loader: load_csv_like_price_series
What you get
- daily gold prices
- multi-currency series when available
Good for
- gold shock studies
- safe-haven behavior
- attention versus commodity price comparisons
Notes
- For tutorials, a CSV download or mirrored source is often the simplest path.