The web portal "Analysetool handhaving maximum snelheid" (or "HotSpots" in short) simplifies the process of exploring locations in need of speed surveillance on the Dutch motorways and helps monitoring the effects of speed enforcement measures. The content presented in the web portal is created on the fly by a Matlab webservice that runs on on a secure server. It does so at great speed: an analysis of up to 3 years worth of traffic- and accident data is completed in less than 5 seconds.
The tool presents the car and truck accidents projected on the Dutch motorway network. This network is partitioned in sections of up to 5 km for the occasion. The interface filters these 5km sections based on criteria such as number of accidents, time-of-day period, maximum speed, percentage of drivers exceeding the speed limit and intensity/capacity ratio. By clicking on a road section key indicators, such as day profiles with travel times, speeding and number of accidents across the road section, are shown in graphs. These indicators can be used to select candidate road sections for speed surveillance and to analyse the effects of them.
The underlying data include road accidents during a 3 year period, the per minute traffic intensities and speeds from over 12.000 locations in this period, and the (dynamic) speed limits on the Dutch motorway network. Left alone security issues, the amount of data and the complexity of the calculations to generate the key indicators make it impossible to build the entire application in a web page and we had to split the application into two parts:
- The frontend with the user interface and the presentation of the data;
- The backend with all the data and data processing.
The web application was built with Angular4 and the components of the Angular Material suite were used for the graphical interface. The Google Maps library was used for the map and the Chart.js library was used to generate the graphs with the key indicators per road section.
The backend consists of compiled Matlab code. Matlab was selected because it enables a fast and easy development process and, when optimized properly, results in very fast data handling. The link between frontend and the Matlab code running on the background was provided by the Modelit MatlabServer Toolbox. The Modelit MatlabServer Toolbox has been proven technology for 24/7 applications ever since 2008. At that time it was developed as part of the online route planner tripcast.nl.
From a technical perspective Hotspots has been an exciting project to work on as various innovative techniques come together in this project. In the process we created many re-usable components. We consider Hotspots to be a template for future web applications, especially if these require numerical analysis and involve large datasets with a geographical component.
Try the hotspots application yourself or visit our video demo (in Dutch).