Datacenter Tracker

A sourced database of US data center construction projects.

A command-line tool that collects US data center construction projects from news articles, SEC filings and grid-connection queue exports. Every stored value records which source supports it and how strongly. The same SQLite dataset can be read from the CLI, a full-screen terminal interface, or a read-only web console.

Open the console

The web console's project table with operator, phase, capacity and progress tracks — The console: every project, sortable, with how much of it is filled and how far each track has got.

A project page showing its figures, each with what it rests on — One project. Each figure says what it rests on — quoted, derived, or missing.

The full-screen terminal interface's overview pane — tracker tui: how complete each field is, and what is holding capacity back.

Terminal output of tracker show, with a confidence explanation and capacity blocks — tracker show: the row, why its confidence is what it is, and where each track stands.

How it works

01
The evidence check runs in code, not in the prompt. The model must quote the article for each value it extracts, and a value is dropped unless that quote actually appears on the page it came from.
02
Every value is labelled by what it rests on: reported, derived, unconfirmed, inferred or defaulted. Values that cannot be confirmed are kept and flagged rather than deleted.
03
Confidence (0–3) is recomputed, never stored. Independence is counted by website domain, so a single source can never reach 3.
04
Progress is five separate tracks: site control, permits, power, construction and commercial. Grid power is never assumed from the other four, because a finished building still waiting for power is the most useful signal in the data.
05
Only the CLI writes. The web console opens the database read-only, and every judgement is computed on the server so the browser and the CLI cannot disagree.

Built with

Python
SQLite (WAL)
Textual
DeepSeek / Ollama
Playwright
SEC EDGAR
US Census
pytest

Status

Running and in daily use. Discovery finds new articles more easily than older projects, so coverage is still being filled in.

Worth knowing

3,294 tests, passing on a fresh checkout with no API key and no network.
Repairing lightly edited quotes raised quote acceptance from 75% to 95%, with no false positives against unrelated articles.