pulseboard

KO | EN

gitlite — search

#hacktoberfest #typescript #llm #golang #go #deep-learning #macos #java #nodejs #windows #vue #pytorch #ios #open-source

(★ 1)

A healthtech catalyst tracker

pulseboard Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Pulseboard

A small healthtech market-intelligence dashboard. It pulls three public data
streams into one SQLite-backed "catalyst feed" for a watchlist of digital
health companies:

* **SEC EDGAR** — recent filings (8-K, 10-Q, 10-K, S-1...) for each watchlist ticker
* **ClinicalTrials.gov** — recently updated studies for tracked topics (digital
  therapeutics, remote monitoring, AI diagnosis)
* **Funding news** — keyword-filtered items from healthtech RSS feeds

I built it because following early-stage healthtech means watching three
different websites that never talk to each other. Now it's one page.

## Run it

```bash
pip install -r requirements.txt

# tell SEC who you are (they ask API users to identify themselves)
export PULSEBOARD_CONTACT="you@example.com"

python -m scripts.refresh          # pull live data into pulseboard.db
uvicorn pulseboard.app:app         # serve the dashboard on :8000
```

No API keys needed — all three sources are public.

Offline demo (bundled sample data, no network):

```bash
PULSEBOARD_OFFLINE=1 python -m scripts.refresh
PULSEBOARD_OFFLINE=1 uvicorn pulseboard.app:app
```

## Tests

```bash
python -m pytest
```

Parsing is separated from fetching, so the test suite runs entirely offline
against fixtures in `tests/fixtures/`.

## Design notes

* One normalised event shape across all sources, so merging into a single
  feed is trivial (see `pulseboard/store.py`).
* Sources fail independently — a dead RSS feed doesn't kill a refresh.
* Dedupe happens in SQLite (`UNIQUE(source, ref)`), so refreshes are
  idempotent and cheap to run on a schedule.
* No frontend framework: server-rendered Jinja2 + a little CSS. For a
  read-only dashboard that's simpler to run and easier to reason about.

See [docs/WALKTHROUGH.md](docs/WALKTHROUGH.md) for a guided tour of the code.

## Roadmap

- [ ] Price/volume context for watchlist tickers
- [ ] Email/Slack digest of new catalysts
- [ ] Per-company pages joining filings + trials + news

## Credits

Built by Euan Brown, developed with AI pair-programming assistance; architecture,
scope and every line reviewed and owned by me. Data courtesy of SEC EDGAR and
ClinicalTrials.gov public APIs.

pulseboard

pulseboard Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

pulseboard

pulseboard Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README