Tuesday, October 6, 2015

Addressing Geography

Doing some analysis of the range of data types that might be important to a new project.

One potential data element is address and there is a lot to be said for treating an address as an entity as indeed it is by most line of business apps. Addresses have a loose structure that is none the less readily understood by humans and within that structure there are data elements that could be extracted and used as feed for a range of analytical processes. 

I recalled an excellent blog post entitled “Falsehoods Programmers Believe About Addresses”  that served as a dreadful warning to anyone intending to walk down an address related path without help.

I also remembered the excellent Open Address project https://alpha.openaddressesuk.org to which I had made some minor data contributions in the past and had looked forward to working with again should opportunity arise. Their public API provided resources for extracting addresses from text and even validating addresses from their growing database. Sadly, this project has been mothballed https://alpha.openaddressesuk.org/blog/2015/07/27/a-time-for-going-to-bed as it looks like their grant funding has come to an end.

This is a not infrequent ending for grant funded research and development projects (as we know ourselves). Expertise, skills and tools are assembled to solve problems – and are then left unused. This particular project is particularly sad as it had the potential to provide tangible benefits to any organisations or businesses anywhere in the UK – not to mention us app developers. It would be nice to think that a new funding model could be developed to get this project rolling again and to provide long term security for the data.

In the meantime it looks like DIY in some form is the safest bet – particularly if turns out to be a good idea to link the results into the geographic expertise we developed for our Sea Kayak Wales project (also mothballed – but for completely different reasons).

I post more technical items here http://mikeoncode.blogspot.co.uk/ and any tricks or techniques discovered in the “address” space worth sharing will be posted there.