With the help of anonymized movement data from mobile devices, important traffic junctions were determined on the basis of three specific events in the Upper Austria region. This should help to avoid traffic jams and traffic chaos at large events in advance.
Existing transportation infrastructure often reaches its capacity limits during events, concerts and functions. To better plan arrival routes and visitor flows (arrival time), Triply offers event planners a web platform that uses historical booking data and attendee numbers to perform forecasts and provide alternative mobility solutions for transportation. Floating Phone Data (FPD) can be used to improve the forecast.
Floating Phone Data is movement data generated by the dial-up of mobile devices (e.g. smartphones) in radio cells. This data is anonymized and can be used, for example, to generate a demand model (source destination matrix) or to support traffic situation services. The mobile phone provider Drei offers a service called Motion Insights, which on the one hand provides ready-made analyses for the areas of traffic, tourism, retail and events and on the other hand offers Austria-wide FP data for own analyses.   .
As part of a customer project with Triply GmbH in 2020, the Smart Mobility and Analytics team of RISC Software GmbH investigated the suitability of FPD for festival and mobility planning using various data engineering processes. The FPD were provided by the mobile network provider Drei. In a first step, an exploratory data analysis and visualization was created to determine the potential and possibilities of FPD. Mathematical and algorithmic methods were then used to build a model using the source target matrix and the temporal visitor flows. The model takes into account the bias of average days when no events take place. The following questions are answered in the context of festivals:
- From which regions do the visitors come, which routes are chosen and where do the routes of as many visitors as possible overlap.
- When do visitors from different regions leave, when do they pass the hotspots and when do they arrive at the event.
This makes it possible to identify those hotspots where particularly large numbers of visitors converge in order to offer special event stops at these locations. In addition, some routes could be coordinated to avoid congestion on the way to and at the event. For the preparation of the models, the data of the following three events were analyzed:
- Clam Rock Festival 2019 (Clam Castle on June 28, 2019),
- Rock im Dorf Festival 2019 (Klaus reservoir on July 4-7, 2019) and
- Kronefest 2019 (Linz on August 22-24, 2019).
Assuming that visitors mainly travel by car and not by public transport, the floating phone data of Drei, which includes the postal codes of the starting points and departure time, was mapped onto a traffic graph. The routes and hotspots identified in this process for the Burg Clam event example are shown in the figure below. The white dots on the map are the starting points of the routes, the blue circles shown in different sizes are those hotspots where the arrival routes of the visitors come together. The size of the circles depends on the number of visitors passing a hotspot. QGIS was used for the visualization.
For the second question, the data from Drei for the festival day were analyzed with a temporal resolution of 30 minutes. This allows statements to be made about how private transport routes must be planned so that they are at the right place at the right time. The model created is also suitable for analyzing daily traffic or company mobility. In the public transport network, it can be used to capture the actual effects of changes, which are often very complex and frequently estimated on the basis of incomplete data. Effects of changes such as road works, additional stops or accessibility can be visualized to support decisions with a good data basis. In the area of company mobility, the routes of the company’s own employees can be superimposed, allowing better and more sustainable mobility to be planned.
Figure 1: Visualization of intersections / hotspots on the way to Clam Castle.