Last week MOJ Digital Services took part in the Public Sector Hackathon in London, and came second for an API!
A ‘hackathon’ is an event in which developers get together - often over a long day or two - to quickly create software from scratch. It’s all above board, and nothing to do with illegal hacking!
The Public Sector Hackathon, organised by MongoDB, attracted about 80 people from across government.
We came second in a competition at the event by creating an application programming interface (API) for data on river levels in the UK.
The Environment Agency holds about 4,500,000,000 hydrology records on river levels, flows, rainfall, etc - with over 100,000 being added every day.
This information is vitally important for people at risk from flooding, climate scientists and others, but was previously hard to get to.
You can now access this data with the new API, which can compare today’s river levels and this year’s floods with data from the past 50 years.
The API gives historic and real time data. You can use it to search for stations and data by postcode, and for specific dates and times.
Our team was made up of:
- Danny Hawkins and me (MOJ Digital Services)
- James Procter and Simon Wood (Environment Agency)
- Jack Harrison (Ordnance Survey)
- Vincenzo Selvaggio, Alessio Gottardo and Alessio Checcucci (private sector developers)
We built the prototype using Ruby On Rails, MongoDB, Grape, Grape Swagger and SwaggerUI.
For the real time information, the system reads hydrology data files every 15 minutes, parses and loads the data into a collection, and exposes the data through an API with a user interface for auto-generated documentation.
MOJ Digital Services will be holding our own hackathon at our HQ near St James's Park in London in the summer - watch this space for details!
3 comments
Comment by Owen Boswarva (@owenboswarva) posted on
The "hydrology data files" link above is wrong. It should be: http://flooddata.alphagov.co.uk/
Comment by Ryan Jarrett posted on
Thanks Owen - that's been corrected now.
Comment by Chris Hansen posted on
This is a really nice implementation! I'm going to take a closer look into some of these techs now that you have shown the way.
Has anyone looked more into this since or thought about carrying it forward? I worked with this data pretty intensively for a week or so and found that the vanilla data needed quite a bit of 'help' to make it useful or, more accurately, less misleading.
I was doing my crunching in good old Python to make ArcGIS Online Map Services.