Open Data Day Cambridge | Software Sustainability Institute

"Open Data Day (ODD) is an international event that runs on the first Saturday of March, started in 2010 and supported by Open Knowledge International. It aims to raise the profile of all types of open data, from government to research. Creating our own ODD The Open Data Day organiser’s guide recommended picking a focus. We didn’t have a huge amount of time to organise, and we knew this wasn’t going to be a large event but mainly a motivation to meet some busy friends. So we kept our ambitions small and meaningful. For Yo and Naomi, the main motivation was to use the opportunity to learn something new together (specifically, Python). So that’s exactly what we made our aim: be a welcoming space to learn something new, and learn together. We co-organised openly via GitHub: Yo made the website and secured sponsorship, Naomi arranged the venue and catering, and together we created the content for our site and the Open Data Day in Cambridge. On the day Open Data Day was the Saturday after Storm Emma, which caused travel chaos in the UK. A few sign-ups sensibly dropped out due to the snowy conditions and in the end we were a small group: Rachel Spicer and Peter Murray-Rust arrived in the morning, and Josh Heimbach joined us for the afternoon. First, we decided what to work on. Peter was keen to use some WikiData, Rachel just wanted to learn and possibly use Kaggle datasets, Yo had intended to make a web app of traffic data, to build on and update a previous project, and Naomi wanted to make her first Jupyter notebook. We settled on investigating bicycle theft using data openly available from UK police. Several in the room had fallen victim to bike theft, Peter very recently, and we were curious to map the data to see if there were location hotspots, and understand patterns of reports by month and police outcome. To begin the task, we downloaded CSVs of all crime reports in Cambridgeshire constabulary, and reviewed the data as a group before taking on different wrangling and visualisation tasks. Since the data was provided as separate CSVs per month, and had columns and crime types that we didn’t need, Rachel combined the files and stripped out the extraneous bits...."