Last weekend (Friday 21st-Sunday 23rd Mar 2014) I went to the second Open Glasgow hackathon, which was based on the theme of Energy. Every one of the four events carries a £20,000 prize for the winning team to develop their idea to commercial application.
I have a fair bit of past experience of working in the subject so was particularly interested to see what ideas came out of it but a five-figure grant to implement an idea definitely helped!
The structure was a bit different from what I was expecting; the organisers seem to be going through a process to find a structure that works best.
On Friday evening, everyone was expected to turn up having thought of an idea and go through a process of “speed dating” it: explaining your idea to another person in a fixed amount of time, listening to them explain their idea and negotiating to assign points to the two ideas up to a combined maximum of seven points. After five iterations of this, if your idea had at least twenty points you had the chance to pitch your idea to the crowd and they had the option of joining a team to work on your idea. Ideas which recruited a team smaller than four were culled.
This filtered out some of the poorly considered ideas but the surviving concepts seemed a bit homogeneous — of the nine surviving concepts the majority were variations on a theme of using the provided data1 on electricity consumption and applying some analysis, which was a bit disappointing — I’d heard more different ideas pitched but they hadn’t survived the process. Speaking to some of the other participants, they agreed with this also.
The idea I pitched, titled Happy Homes (or Happy Buildings), fitted into the common theme too: trying to put a face2 to energy consumption with the theory being that this might be more accessible than the standard graphs and numbers.
After working out “normal” electricity usage for Glasgow City Council buildings this could be compared to current use, with “abnormal” usage represented with either a sad (above normal) or happy (below normal) face. From that point it would be trivial to replace the provided data with input data from a Current Cost or smart meter, for a more domestic application.
It survived both stages of the pitching process and five of us built a working prototype, plotting a map of the locations available, simulating the current power consumption and doing the comparison to “normality” to decide which face to show.
This was done using Node.js to construct an API, backed by MySQL, and a web front-end (using the Hackathon Starter) which plotted the locations on a map. One of the team members built an Android app which showed a similar visualisation.
We didn’t win but it was still a good weekend. The winning concept appeared broadly similar but focussed more on the idea of encouraging action — notifications would be sent out to responsible parties (caretakers, etc.) if consumption fell outside “normal” so that they could check that, for example, lights hadn’t been left on.
As far as future development goes, one significant functional requirement would be getting automated data ingestion working. I’d experimented with adding D3.js-based graphs on the Sunday afternoon but didn’t have enough time to get them integrated with the Jade templates the front end used.
I’m also inclined to re-implement the server-side components in Rails — Node.js was good enough for getting a working prototype built in the short time available but I want to experiment with Resque or delayed_job.