On Saturday the 21st of March, I participated in “Apps for Ghent”.
Along with its digital partners, the city of Ghent organized this hackathon – completely in the spirit of the ‘smart cities’ evolution – in order to bring open data enthusiasts, developers and actively participating civilians together in a co-creation event. As I feel that I have traits of all three, an obvious match was made and I planned that rainy early spring Saturday to be “hacking”. In order to achieve anything over the course of this Saturday, I had to team up with other data-eager participants. Luckily, I found two LoQutus colleagues that were as excited as I to participate in the event that focused on possible new and insightful uses for the data of the library of the city of Ghent. Accordingly, the event was called ‘Hack the Library’.
For the first time, the data of the library was open to the public. Two weeks before the event, a ‘datadive’ was organized where the data sets intended to be released were presented. The data were related to over 35 million (anonymized) book loans, over 250.000 members reading over 500.000 books. The data itself would only be available the day before the event. To make things even more exciting, the driving force behind the open data initiative in Ghent, Bart Rosseau, shared with us that we would need more than our basic home, garden and kitchen resources in order to tackle this problem. Needless to say that the pressure was on.
We, my LoQutus colleagues and I, decided to develop a platform to quickly extract insights from the provided data. We built this platform using the open-source statistical programming language R and the visualization tool Qlik Sense. It allows users to navigate and explore the extensive data set themselves and to produce insightful dashboards. Such a self-service BI (Business Intelligence) environment adds great value to the use of data to support fact-based discussions, which might steer a business (in this case the library of the city of Ghent) into a new direction. I especially believe that in the context of a ‘smarter city’, such environments might become increasingly more popular, since they invoke discussions and allow users to support their opinions with clear and beautifully visualized data.
More in detail, we achieved the following during the six hours hacketon:
- We wrote code to quickly handle data cleaning steps;
- We combined our ‘cleaning’ code with user documentation;
- We built some very intuitive dashboards;
- Which ultimately delivered a reproducible analysis (i.e. added value for the library and its users).
By the end of the event, this allowed us to visualize a membership dashboard with the membership number evolution between 1996 and 2015, the members per statistical sectors of Ghent and the members per birth decennium. In addition, we created a book dashboard, which shows how the loans for a particular book evolved over the years, the most popular books for a selected author and a milestone spike in 2008 for the lending and reading of ‘Het Verdriet van België’, when the author Hugo Claus deceased. Not without a sense of shame, I have to admit that I have not read this masterpiece myself.
I am a great supporter of the open spirit of smart city innovation in which the Apps For Ghent initiative was conceived. Therefore, my LoQutus colleagues and I decided to reveal our data cleaning and the Qlik Sense configuration. You can find the data cleaning on Rpubs and the Qlik Sense configuration on Slideshare. I have to admit to anyone who looks into the proceedings of our participation that a six hour hackathon does not lead to a very polished result. However, as a take-away from the experience, I am confident to propose a similar approach for any data initiative, if you have raw data that you want to mine or if you want to set up a sandbox environment for quick extraction of value and insights from this data.
For more information, do not hesitate to contact us!