Development and Evaluation of Tools for the Investigation of Temporal Patterns of Traffic Accident Data.

The aim of this bachelor thesis was to develop a dashboard that allows the user to analyse temporal patterns and structures of traffic accident data using different filters. The results of the visualisations are then be evaluated, with a focus on the possible detection of temporal peculiarities of the data.

The data from the accident atlas of the Statistische Ämter des Bundes und der Länder served as the basis for the data. These containe information on traffic accidents in Germany. For this work, the data was limited to the region around Munich and Augsburg and the years 2016 to 2021. The data was processed in JupyterNotebooks and displayed using dynamic Plotly visualisations. The dashboard was developed with Dash, for a better usability for the user. In the finished dashboard, the user can use filters for individual years, compare two years or get an overview of annual developments.

A large part of the work then deals with the results of the individual visualisations regarding the question of whether certain temporal patterns of the data are recognisable. Above all, time-of-day structures could be detected, as well as seasonalities of individual road users. In addition, the comparisons of the data before the COVID-19 pandemic with the values from 2020 were particularly striking in terms of how the restrictions have affected the traffic accident situation.

Github: https://github.com/Leonieen/BA