Disruptive Proactivity

The school day may end at 3:45, but children don’t simply vanish until the next morning. They may go home, or to after-school activities. The choices available are important.

There is no central source for those activities in schools, and collecting it would be disproportionately difficult going school by school (and would use a lot of teacher time). Those activities may be supported by external grants from BIG Lottery and others, who do therefore have a national view, if only partially so. BIG receive requests for support, and know what they fund. Transparency on where they put resources in this area gives a key insight into provision.

A little while ago, BIG published their grant spending for the last few years; a set of big spreadsheets, which are here tidied up. At the request of the nice (and very patient) people at the Indigo Trust, I had a look at what the BIG grant data could do to start to help with that question, and why it is more generally important.

Why is this data important?

There are a number of issues that matter in practice, that are almost impossible to measure directly. But their presence, or absence, can be inferred from their effect on the world. We know planets orbit around other stars by detecting the reduction in light that occurs when they pass in between it and us. The BIG Lottery Fund “money goes to community groups and projects that improve health, education and the environment”. Quite broad, but they helpfully have an applicant category of “State School” in most years for which they have published data. In total, we have 13574 state school applicants.

The BIG funding data is vitally important, both because of their profile, and the focus on easy access to smaller awards. Those awards are for the type of projects that only happen because small pots of money appear from external sources. They may be small to most funders, but they are larger than can come out of a regular PTA budget. We can detect possibly different provision by examining the requests that pass in front of BIG’s review process.

We won’t know from their data why schools have never received BIG funding for activities (otherwise funded? nonexistent? turned down?), this is only one piece of the puzzle.

This work does not seek to suggest causality, merely looking for correlations for further study as a proof of concept of what could be done.

Does it matter how long it takes to get home?

The lovely people at PLACR have a new API which allows you to take a lat/long, and find the nearest public transport (of whatever kind you choose). This small tool should very quickly form a vital part of the transport API infrastructure.

We took the school names and local authority from the published spreadsheet, ran them (very slowly) through a geocoder, to get a postcode, and from there, got the nearest train station (thanks again to PLACR). With the nearest train station, we can ask trainTimes.org.uk what time someone would get home, if they left London at 5:30.

It’s imperfect; Norfolk doesn’t really have any railways, so you can’t get the train anywhere, and the closest station may be a branch line with a major station with trains fast to London more effective but slightly further away.

So does it matter how long it takes to get home? The answer is not really. Any effect comes from something else.

If you’re in Scotland, you’re more likely to have been successful in high counts of years but we get no idea why.

That data is available here, grouped by award and by school (or dropbox). There are many many questions that can be asked from these columns. I’ve only looked at one of them so far.

Need for this data

None of the above was designed to be a rigorous and statistically significant project. That wasn’t what they were for. Those three give some data driven insight into interesting questions, about which data does not exist. BIG doesn’t have a complete dataset, it wasn’t what it was creating.

BIG, releasing only the data it has, is a huge step that makes something imperfect that was impossible before.

Releasing the parts of data that one organisation has, is enough for someone else to put together something helpful for other pieces. It is unlikely that BIG will do an indepth piece of research on travel times around their school grantees, but some better regression analysis (controlling for the different types of school etc) would probably be worthwhile to know. It would be impossible to begin to ask the question without that data. Linking schools to ofsted data and doing some keyword/text mining on the “Press Summary” field would be very interesting.

BIG can’t publish all historical addresses, as some of them include home addresses of individuals (for small organisations), and that can only be identified by a change to the form for future requests. But where organisations have charity numbers, company numbers, and other forms of identifier (but unfortunately not, DfE school identifiers), BIG have included them. That allows linking with other sources and the asking of novel questions, of no use or interest to the original owner. When multiple

That’s why the open data is important.