Spatial Simulation

Final project: What is the impact of cycling preferences on the spatial pattern of cyclists?

The agent-based model was designed to obtain information about the way how cycling preferences influence to the spatial pattern of cyclists. The model takes into account three main factors that influence on a cyclist's choice. They are: the length of a road (the shortest path), preference of safety road and the slope degree of land surface. The hypothesis of this model is that cyclists prefer safety road even if it is longer and take more time. Also, by modeling the cyclists' preferences it is possible to identify the most frequently used streets and thereby pay more attention to their infrastructure in the future. For validation purposes the ground true data obtained from counting stations in Salzburg city were used.

An overview of modeled entities
An overview of modeled entities

I made 3 experiments. In each of them the cycling preferences were changed. Below I showed the results obtained from these experiments and heat-map to each of them. It helps visually evaluate the results. For each scenario, I put the following general characteristics:

  • - 200 agents;
  • - return to the place of departure;

I made 10 repetitions for each scenario, and used the average value of the received data.

Scenario-1: Cyclists prefer shortest road

Scenario-2: Cyclists prefer safest road

Scenario-3: Safest road + slope index

After exploring the real data obtained from two counting stations I found a certain regularity. The percentage ratio of registered cyclists at two stations is stable and shows around 85% on Rudolfskai station and 15% on Wallnergasse station (see table below). For validation I used the data obtained in 2013. I translated the data obtained from the experiments in percentage ratio of registered cyclists on both stations. Thus, my strategy for the validation is to compare the percentage ratio of registered cyclists in real life and in the model. Scenario which will show relatively similar percentage ratio between gathered data will identify the main cycling preferences in real life.


The hypothesis defined in this model is derived from human’s self-preservation instinct. People always have a safety in a priority. Therefore in this model the hypothesis was that the cyclists choose the safest road even if it is occupied by more time and the road will be longer. After comparing the data obtained from counting stations with the simulated data, I can assume that the hypothesis of this work was confirmed. Cyclist tend to choose safest road even if takes more time.