This repository contains the code that generated a prototype that our team presented at the datathon 2020 organized by the faculty of science of the KU Leuven. There were 40 teams participating in the datathon. Our team managed to go home with the prize for best presentation.
The KU Leuven Datathon is a datamining competition in which teams of KU Leuven Master and PhD students work to discover and share insights regarding a large, rich, and complex data sets. The event is organized by the KU Leuven Big Data Platform and in collaboration with the Smart City Project of Leuven. During the edition of 2020, the students were provided a dataset that contained measurements of traffic and weather data within the city of Leuven. There was no research question asked. The goal was to do something valuable with the data.
The data from our database was collected using sensors. When you collect a large amount of data froms ensors, it is vital to monitor the quality of the data. An aspect of this quality monitoring is outlier detection and then deciding what you want to do with once you've identified these outliers. We created a dashboard that allows the user to quickly identify outliers. We then provided the user with data of news events and events organized by the KU Leuven. Using these data sources, the user of the dashboard can decide whether the outlier is a measurement error or if the data point is because of a certain event that took place. The jury applauded our solution for our creativity and the depth of our solution. Our team was awarded the prize for best presentation
Ruben Kerkhofs, Mitja Briscik, Zachary Jones, Francisco Javier Jara Avila,Jack Xia