Mobility Data Analysis (Modul MW2436)

See TUMonline

After having participated in the course the students will be able to name the essential elements of a data analysis pipeline for mobility data and build a corresponding framework with open source software. In addition, the students are familiar with different possibilities and formats of data collection, aggregation and storage. They will be able to record mobility data and evaluate the collected data. In addition to classical statistical evaluation methods, they are familiar with other methods that are especially relevant for mobility data, such as hotspot analysis, spatial clustering, geo-fencing and simple machine learning methods for classifying modes of transport. By applying these methods, students can question the collected data critically using domain-specific indicators and generate corresponding visualizations.

• Fundamentals of (Geo-)Data Analysis • OSM/GIS • Basic Methods and Visualization • Spatio-temporal Data and Clustering/Heatmaps • Experiment 1: Data Acquisition and Driving Behavior • Experiment 2: Personal Mobility Analysis • Machine Learning Classification Project

Python Basic Knowledge

The module takes the form of a practical training. A appointment takes place in 1-2 blocks depending on the date. Each block starts with a theoretical introduction delivered as frontal knowledge transfer by means of presentation and live programming. The students then work on concrete practical tasks in the form of supervised single and group work. In two sessions, experiments on mobility data analysis are performed, with the students actively participating in the experiments under supervision. With these methods the students learn, to know the essential elements of a data analysis pipeline for mobility data and are able to set up a corresponding framework with open-source software themselves. They will get to know the different possibilities and formats of data collection, aggregation and storage and will be able to carry out mobility recordings independently and preprocess and evaluate the collected data.

The module examination consists of written short tests (answering short questions, 25 points, 10 minutes, no aids allowed), homework (programming project, 25 points, all aids allowed) and a short final project (90 points). The examination of factual and detailed knowledge and its application (based on the exercises) is held in the beginning of the following appointment either in the form of a test or a homework assignment (being announced in time). In this way, the students demonstrate that they know and can apply the essential elements of a data analysis pipeline and GIS. After the last date, there is a graded final project (working on a given programming task). In this way, the students demonstrate that they can independently apply the evaluation methods and use simple machine learning methods for classification and evaluate the collected data. The overall grade calculated of the sum of points from three short tests, three evaluated homework assignments and the final project.

Thomas A. Runkler, Data Mining: Methoden und Algorithmen intelligenter Datenanalyse, Vieweg+Teubner Verlag (2010) Baoguo Yang, Yang Zhang in Advanced Data Mining and Applications (2010)