Both sides of the COINS

Last Friday saw the coalition government publish the Treasury’s COINS database for the first time. That might not sound like much, but to data geeks like me it was very exciting news; so exciting that to my shame I woke up very early that morning, like a child on Christmas Day.

The COINS database had acquired semi-mythical status amongst those who want government to share more of its data – see the Guardian’s freeourdata campaign. The database – its full name is the "Combined Online Information System" – was billed as a comprehensive record of government spending, showing exactly where and how the state spends taxpayer’s money.

The NCVO research team is very interested in the financial relationship between the state and the voluntary sector (so interested we released a report with that title). My excitement about COINS was based on its potential as another source of data on this financial relationship.

We know from data from charity accounts how much the sector says it gets from government, data like COINS complements this by telling us how much the government says it gives to the sector. We have the Office of the Third Sector’s surveys of central government departments (sadly no longer online - a casualty of the name change I think), but access to the raw data would allow much deeper analysis of the relationship, and would serve as a check for the charity account data.

So on Friday I got to unwrap the Christmas present, and unfortunately since then I’ve been slightly disappointed.

1. Firstly, it’s a very unwieldy database, consisting of millions of records across tens of fields. It took a long time to get into a useful form (an SQL database) and for technical reasons was quite difficult to manipulate.
2. Secondly, while the database was seen by some as a treasure-trove of data, HM Treasury officials seemed to be keen to play down its significance, and they were probably right. It does contain millions of records, but those records are separated not by supplier but by project. This means we can’t easily match source with recipient; we have to guess which projects we expect to fund the voluntary sector.
3. Thirdly, the language and codes used in the database are often impenetrable to those without specialist knowledge. The data community are working hard to find these out – either through diligent searching or through freedom of information requests.

But it’s not all doom and gloom – this represents significant progress towards open government data and with work it may provide useful data for the voluntary sector. We’ll be keeping an eye on similar developments over the next few months, focussing on what they can add to our knowledge about the sector.

Find out more about COINS:

(Oh, and apologies for the pun in the title.)