The database

Check any figure on this site

Everything this project computes comes out of one SQLite file, and that file is public. You can browse it table by table, query it in the page, ask it a question over the web with no key and no account, or download the whole thing.

It is not an official record and it is not the town’s system. The source of truth is a set of CSVs read out of documents the town, the district and the state published; the database is a derived read model, rebuilt from scratch every time, and nothing is ever edited inside it — a row in a database has no address, no publisher’s filename and no sha256, and those are what make a figure checkable.

So the useful claim is narrow, and it is the one this area exists for: every number we publish can be traced back to a document somebody else published, and you can do that yourself without asking us.

Start here

Written by this project, not by the town or the district. Every figure on these pages is computed from sources/data/lunenburg.db or from the documents in the archive, and a page with no generator listed is hand-written — its figures were true when it was written and nothing checks that they still are.

The database, table by table — Lunenburg Budget Project →html · 424 KBEvery table in the database, what it is for, which years it actually covers — and, first, whether the question you arrived with can be answered at all.regenerated by scripts/build_schema_page.py

What can and cannot be joined

The school money graph — every node, inputs and outputs →Every node in the school money graph as a flat list, inputs and outputs, each marked traced, partial or unknown. The rows with no number are the point.The Lunenburg Join Map →Six levels of detail, the key each one turns on, and which of them the district’s budget and the town’s books can actually be brought together at.Can I Match It? →Three levels of the school budget against four things to match them to. Each cell says what is true today and, separately, what would open it.Lineage Graph →Question to key to table to extract to report, as a graph — and the three different ways a connection fails, only one of which a schema can show.

If you would rather have the file

Read-only, open to anybody, no key. The two discovery documents describe the rest, including the four ways to get a confident wrong answer out of this data.

/api/indexWhat exists, as JSON. The map a program should fetch first./api/schemaEvery table and view, what each is for, and the four ways to get a confident wrong answer./data/lunenburg.dbThe whole database, SQLite — 19.2 MB. The same file every figure on this site comes from./data/archive-manifest.csvEvery document in the archive: its address, the publisher's own filename, and a sha256.

Every other machine-readable address is listed on .

The documents themselves

This database holds figures we extracted. It does not hold the PDFs, spreadsheets and workbooks they were extracted from — those are the archive, and they are the thing to check when you do not believe a number, because they were written by the town and the district rather than by us.