As someone tweeted, its interesting that on a day thousands decide to protest in central London against the cuts to government services, others get together to try and create solutions.

I attended the third National Hack the Government day, which, contrary to some early reactions from friends and colleagues, is not some shady underground movement where people try and break things, but an energetic session where people with a combination of  coding skill, curiosity and a general wish to make things work better get together and see what they can do with government data. Its run by Rewired State and todays session was at Kings Place – home of the Guardian.

Notes follow on most of the presentations (apologies if I missed you out, or got your name wrong (or just didn’t get your name….) – blame my scribbled note taking or the fact I simply couldn’t get my head around what you had done!) Full details and links to all the material available should be published on the Rewired State website.

Young Rewired state team – iSteal. They used crime data and produced an app which helped you to calculate the likelihood of you pulling off any particular crime. nb uses amazon-type suggestions to advise alternative crimes in an area which have lower detection/capture rates!  Winner of the Evil Genius award.

Tom Morris – Trained to complain. A simple form which auto generates a letter for you to either email, or print, sign and send to rail companies. If we all complained when we experienced delays, perhaps  rail services would improve?

Two presentations followed which showed ways of illustrating how the Alternative voting system would work. One had a triangle, where the red, blue or yellow segments grew or shrunk in size, the second had a target style chart, with the inner ring being first choice, outer ring second choices. Then, when lowest is eliminated, the image shows you how votes are redistributed.

iBill – real-time progress of things happening in government, using an RSS feed. Neat features include allowing comments which are displayed classified by sentiment analysis – eg broadly positive = green on left, negative = red on right. Other ideas include informing people when they had expressed negative statement about any protests which related to that topic, or possibly also integrating with they work for you – for direct comment to those who should be able to react.

Courtfinder 2.0 – based on current HMCS site, but improving search and data display, to provide a much simpler experience,  including google maps.

Fear itself. The idea behind this is that the more fear inducing artcles are published, the more fear we will feel. The site this distills articles from a range of newspapers and produces a daily threat level. nb used python natural language processor to classify stories.

Not get stabbed. The first of several games. This one based on police data and google street view.

Top trumps – another game using crime data, this time centred on London boroughs. You play against the computer – when it wins, it suggests a category, when you do, you keep control.

Another presenter first described his plan to do a Top Trumps style game, this time to find out information about MPs  (Tom TruMPs – get it?!) – but due to problems getting the data he imagined, this remained at concept level.

Nabbd – an app to improve the quality of data collected asbout possessions, with a view to making the process of reporting crime more streamlined and better both for the victim of crime, and those who are involved through the process.  This won the “Most commercially viable” award, given by the Government Technology Strategy Board – so I look forward to seeing it evolve into a real product. The idea is that it should be easy for people to input data about stuff – the examples given were shiny geek toys (yes, he had an iPad 2) but the model would work equally well for bikes, designer stuff etc. You input data on item, including taking a photo – then, if it is stolen, the system makes the process of reporting it to the police straightforward: it provides all the information from the record, generates a mail, and also (I think) a process to track your report.

Should I wear a coat? – for times when looking out the window is not an option! Uses geolocation and combines with data scraped from the met office and google to give you a recommendation.

Map what you like: translates CSV data onto a map – time series and geog data

Looquest. Another game – which also helps crowdsource data once you find one. This presenter deserved an award for being the most entertaining and lively!

Neon Tribe presenter

Neon Tribe presenter. Photo credit: Neil

He also presented Tim Davis’ hack on the proceedings of parliamentary committees. It displays the record in the style of a restoration comedy. Intriguing was the marginalia – each intervention was classified by keywords, wordle-typecloud, sentiment analysis and more.

Usefuliser. The example given was of police data – where many files are produced, per constabulary, per month and it becomes tricky to work with so much basic data. This process cleans up data and adds unique ID to each line of code and/or per record set – makes it much easier to reuse. It too won an award – I think in the Police data category.

Crime map. Again uses police crime data – move pin on map, find out crimes in area.  When you select a specific crime, different colour pins are added to show where they happened.

Dylan presented Where should all the convicts go? I think his idea was that as prisons are overcrowded, you could uses the datasets of numbers  and mash it up with populations of countries to find out where would be the best place to send people – Hmmmmm

Courtsquared. Scrapes HM courts site, and looks at all cases in a particular court. You  can check in and submit your own report if you attended (incl graphics) – to counter fact that most court cases are under reported. This won the “best use of Justice data” award.

Next followed a curious app which somehow coded your daily ‘journey’ in colours – it looked hypnotic, but I really didn’t understand what they had done (sorry guys!)

Crime4U was another hack mapping crime data onto a map.

Harry Wood presented an open street map hack which showed bus routes rendered onto a map – his used routed plotted by the open street map community, but TFL routes could be overlaid to compare.

The London crimebot was presented by a guy who said he had only come to take photos, but thought he might as well get involved and make something. He used crime data as many others did, but the app was designed to be used via twitter – your phone would geolocate you, you send a tweet asking #crimehere ? and the response presents the top line figures in a tweet (not sure if you are in a crime hot spot if it then suggests your quickest route out of there?)

Next was an app which plotted crime data and traffic accidents onto a map.

Then a neat un-named tool which wraps  html around data from spreadsheets and allows simple display with formatting for a slick presentation.

Scandalous – which won the transparency award – aimed to make clear the whole picture of an MPs life – by mashing up data on their activities both in and out of government.

Find me a region put control into your hands, enabling you to find the best place to live by rating according to which criteria were most important to you – for example house prices, crime etc. Ideas to include it included changing the boxes where you had to input figures to slide bars < neat idea, I think it won the “Disruptive ” award (not sure why!)

Another use of crime data focused on re-offending rates and plotted areas with high rates as hot spots on a map. He then compared this to areas with high numbers of probation officers – but actually the surprise result was that more officers = higher rates of re-offending…..  but then he mapped the hot spots against unemployment – a much clearer correlation.

The next hack was highly topical – aiming to combine twitter data with traffic cameras to enable you to display a real time view of the route of a protest. Unfortunately, not enough tweets were geotagged, and spookily, as the protest approached, traffic camera data ceased to be available. However, he did manage to provide a good display of tweets, showing who was saying what, the most popular terms and the most frequent tweeters.

On to the last 5 presentations:

Presentation of an interactive site which was put together to promote the Digital Shoreditch festival – various ways of collecting and sharing data about businesses in a small area.

Crime data and CCTV – which apparently is currently held by individual councils, and this idea includes  a plan to map it nationally.

Not sure of the name of this one, but Rob’s presentation was of a Government Spending CSV data validation service. Update: Live version now available. This tool would report non valid CSV files and check that essential fields are present. Potentially a really useful tool – both for those inside government departments faced with the relentless task of releasing data, and those outside checking that it is done consistently. This idea won the award “Best in Show”.

Presentation of the Best in Show award

Presentation of the Best in Show award. Photo credit: Neil

The next presenter showed a neat tool that checks which websites track  usage via cookies. I think this followed on some earlier work which focused on NHS choices and has already resulted in follow-up.

The final slot was what sounded like “Wo me lo lo”  - and it means “I’m lost”. This uses loads of different categories of data, and you use the geolocation funtion on a smart phone, and it tells you where your nearest bus stop, pub, school etc is.

Phew – a long session – amazing how much work had been done in such a short time, and while I did write notes, I’ve probably got some descriptions wrong – so feel free to comment on this post and put me straight. I definitely didn’t get all the names right – as evidenced by the fact that awards were announced for Smackdown! – as the most aesthetically pleasing app, and a special interworld award for “Think of the Children” – and I dont recognise either!

The hack day included two plugs – one for an event relevant to my own work: a group in the Netherlands are holding am “Open Data for Development” camp – from 12-13 May in Amsterdam. Go to openforchange.info for more, or follow @open_for_change on twitter.

The second is vitally important if the skills and enthusiasm for doing useful and interesting things with public sector information is to continue beyond the current community of activists.  Young Rewired State is planning another week of hacking data this summer – August 2011 – and is looking for new people to take part. If you know someone under 18 who has an interest in coding or programming, please direct them to the site to find out more.

The day ended with a party in a local pub to celebrate Rewired States second birthday.

Rewired State goodies

Rewired State goodies

UPDATE: Steph Grey has also blogged – with some insightful comments on the whole concept of data hack days. Worth reading.

UPDATE 2: Official list of winners – Rewired State blog

About these ads