The first MiraiLab, “From Prototype to Production” took place on Tuesday, September 24th, at Impact Hub, Colab.
MiraiLabs is a series of data science workshops aimed at professionals, and in this first event we have decided to focus on a very common scenario which often is the source of unexpected problems and inefficiencies: moving your analysis into a productive environment.
In the workshop, we focused on the steps that are required for reaching a productive environment, highlighting the best practices and potential bottlenecks of the process. To this end, we walked the audience through a simple test case: an analysis of the delays in Zurich’s public transport network (VBZ) throughout a four weeks period. The data set is publicly available at Zurich Open Data. The analysis was implemented as an R package, featuring R markdown for rendering stand-alone reports as well as a comprehensive set of results, collected as a website deployed to GitHub Pages.
The material of the workshop, including the slides that were presented, is available on GitHub in the MiraiLabs repository, whereas the final outcome of the workshop is available on Mirai’s GitHub as repository vbzdelays) for the constructed R package.
We hope this workshop has helped in understanding the benefits and challenges of a well-structured pipeline to production, and are looking forward to the next MiraiLab.