About
Good civic work is badly indexed. The organisation that already built the dataset you need, the tool that would have saved you three weeks, the archive holding records you assumed were lost — these exist, and finding them mostly depends on knowing someone who already knows. This is an attempt to write that knowledge down.
The name
Soochi comes from the Kannada ಸೂಚಿ (sūci): among its meanings is an index of a book's chapters or subjects. That is the job here too — a small, ordered way into a much larger body of work.
The seven kinds
Every entry is exactly one of these. The vocabulary is deliberately small — an index where each entry invents its own tags is a list, not an index.
- Organisation A body doing the work — research, advocacy, service delivery 8
- Data source Data you can download or query here — one survey or a whole statistical service 50
- Directory A curated list that sends you elsewhere — it holds links, not data 2
- Tool Software you run to do a job 2
- Publication Writing and reporting you read 7
- Dashboard A live view of data someone else keeps current 13
- Archive A record kept so it is not lost 4
- Community People to ask, and a place to ask them 1
Alongside these, 31 topics and 44 places, both closed lists. Adding a term is a deliberate edit someone reviews, which is most of what keeps the filters meaning anything.
What gets listed
In
- Organisations doing civic, research or advocacy work
- Open datasets, and the portals that hold them
- Tools you can run without buying anything
- Reporting and research that shows its sources
- Archives, including of things that no longer exist
- Communities you can actually join and ask questions in
Out
- Consultancies and vendors selling into the sector
- Products whose free tier is really a trial
- Anything primarily promotional about itself
- Work you cannot reach without knowing someone
- Individual reports — we list who published them, not each PDF
Paywalled work is listed when the work itself is the point, and labelled as such. The bar is usefulness, not size or reputation.
How it stays honest
Every link is checked automatically each week. Entries that go dark are marked dormant, then dead, rather than deleted — what an organisation used to be is often worth knowing, and a directory that quietly drops its failures is lying about its own history.
Descriptions are hand-written and capped at 160 characters. Nothing here is scraped. That cap is doing real work: it is the difference between a page you can scan and a page of marketing copy someone else wrote.
The data is the point
Every entry is a plain text file in a public Git repository, validated against a schema. The whole index is available as index.json under CC0 — take it, fork it, build something better with it. Nothing here is locked in a database that only one person can reach, and if this site disappears the index itself does not.
Who maintains it
One person, with help. The index is only as good as the number of people putting things into it — if something is wrong, out of date, or missing, say so. Corrections are as welcome as additions, and get merged faster.