Nestoria Interview - Anna Powell-Smith
Hear Ye, Nestorcommunters! A few weeks ago now it came to our attention that someone had done something new and cool with the api. The person behind that cooless is Anna Powell-Smith, who ran up a train times vs. house prices graph. It turns out that as well as being a coder, writer and data analyst for hire, Anna is a long-standing volunteer for mySociety, the UK's leading developer of democratic and civic websites. She's also worked for the Open Knowledge Foundation, the open data campaigners, and for start-up ScraperWiki. And she made the Domesday Book freely available online for the first time. Which is pretty cool, so this month she's our interviewee. Thanks for sharing your time, Anna, and the rest of you - enjoy!
How would you describe your work and what would you say your main work interests are?
I'm a front-end web developer, running a little agency, with particular interests in civic coding, open data, online mapping and data journalism.
Can you talk us through also how and why you got into coding - I get the impression that there's an interesting story there?
It was Napster! I was at university studying English when I installed Napster and was blown away. I thought being able to share music with people like that was just amazing, it really made the world a better place - not because it was free necessarily, but the technical achievement of P2P and the software. After that I knew I wanted to learn to code. It was a bit of a challenge as I didn't know anything about computers, or anyone who was into them. However I went back and did an MPhil in computer science and slowly taught myself how to be a developer.
Cool! Can you talk us through the why and how of your recent train times vs. house prices graph? Did it reveal what you thought it would, or were there some surprises?
My partner and I started house-hunting on the basis that house prices would get dramatically cheaper an hour away from London, outside the commuter belt. But then we realised we didn't know if that was true, so I decided to use Nestoria to find out, because it has the best API of any house-hunting site. Working with the API was really a pleasure. It was quite satisfying to see from the graph that house prices do start to fall off more steeply once you get to 60 minutes from London, pretty much what we'd predicted. If you visit London a couple of times a week, the sweet spot looks to be around 90 minutes - too far for even the most hardcore daily commuters.
I was surprised by the bump in house prices at around 4.5 hours from London - it definitely shows that Edinburgh has its own economy!
Google tells me that you're very involved in open data projects - can you talk about some of the projects you've been involved in recently - particularly ScraperWiki, which sounds very cool.
I've worked with the Open Knowledge Foundation, who are brilliant campaigners for open data and a rapidly growing international movement. I worked on their OpenSpending project, which aims to track government spending all around the world, and has just been awarded funding by the Knight Foundation.
And I worked on the early days of ScraperWiki, which is a start-up building a community of journalists and coders freeing up data.
Can you also talk us through the basic genesis of ScraperWiki. And give us some examples of some revealing scrapers?
ScraperWiki was set up by Julian Todd, who is a one-man coding powerhouse crossed with an investigative journalist. He set up PublicWhip, which scrapes the voting records of the UK Parliament and turns them into structured open data. So Julian had the original idea of a code wiki, and Francis Irving, the CEO, has turned ScraperWiki into a fully-fledged start-up. Coders like it because they can share and run code easily, and journalists like it because it helps them find data.
One story ScraperWiki exposed was about corporate sponsorship of all-party groups in Parliament - a coder found that corporations and interest groups channeled more than £1.6 million to MPs and lords in one year by sponsoring these groups, and the Guardian reported it.
Are there any datasets that you'd love to have better access to or that you've struggled to find and would like to work with?
Oh, so many! I'm interested in land ownership, and am regularly frustrated by the Land Registry's ownership records not being open. And about half the land in Britain is not actually registered with the Land Registry at all. That's a really important dataset for transparency.
Fares data for trains would be great, too. And as a house-buyer, I'd love to get proper comparisons between asking prices and selling prices, although it would be a hard one to do properly. One for Nestoria, maybe?
My dream dataset is Nikolaus Pevsner's guides to architecture. They're in copyright, but I'd love to build a location-aware app to show you what Pevsner wrote about the buildings nearby.
Thanks for the property related suggestion - I'll put to 'the people'm the architecture app sounds really cool - I'd definitely use that. Do you see an opening up of data availability or do you see privacy concerns shutting off information further?
Wow, that's a big question! In the UK, there are some great data releases going on at the moment - from live bus timetable information to Land Registry house sale prices. And there are lots of interesting start-ups being built on open data, like Nestoria, and OpenCorporates. So in the short term it's quite positive, but I think the open data movement needs to focus just as much on the long term, because the political environment could always change. We need to make sure we get proper governance and legal protection for open data.
And one more thing - have you come to any conclusions about where you're going to move to yet?
Heh! Yes, the scraping exercise definitely helped a lot - we like Stroud, Sherborne, Castle Cary and Macclesfield. Now I just have to work out how to use Nestoria to find the perfect house...
Here's an old rectory in Stroud that looks nice...
If you want to hear more of Anna's thoughts you could follow her on Twitter.
