Data analytics and road traffic go hand in hand.

A project fused data science with transport logistics in Sweden. It is a collaboration between RISE, Volvo cars & trucks, Scania trucks, and the road traffic authority, Trafikverket. It ran between 2013-2016.

It was financed by Vinnova, a state funding agency. Some information (in Swedish) is available from FFI Vinnova.

The project uses data processing systems running at high rates to produce value to the manufacturers and road operating body from the data they collect. The use cases we have worked with are:

1. Road traffic accident data logged by the police
2. Inform fleets of vehicles during a hazard warning press
3. Detecting road queue accumulation within Stockholm

Business models

Big Automotive Data Analytics, Per-Olof Svensk, Magnus Andersson, Trafikverket and RISE Viktoria, Slides

Business models: State-of-the-art, Abstract

Business Models: Future Scenarios, Report

Big Automotive Data Analytics: State of the art report, Fehmi Ben Abdesslem, Report

Big Automotive Data Analytics: Scenarios, Magnus Andersson and Maria Schnurr, RISE Viktoria, Report

Final Report

Papers and reports

1. Data Readiness, Björn Bjurling, Per Krueger, Ian Marsh, RISE SICS, Report, Slides
   
2. Algorithms for data mining and machine learning, Ian​ ​Marsh,​ ​Bjorn​ ​Bjurling,​ ​Ahmad​ ​al-Shishtawy,​ ​Anders​ ​Holst, RISE SICS, Report, Slides
   
3. Big data architecture (batch processing), Martin​ ​Neumann,​​ ​​Ian​ ​Marsh,​ ​Bjorn​ ​Bjurling,​ ​Ahmad​ ​al-Shishtawy, RISE SICS, Report, Slides
   
4. Audio analytics: State of the art report, Björn Bjurling,  RISE SICS, Report, Slides
   
5. Computer vision survey Applications for the Automotive Industry: A literature review, Abubakrelsedik Karali, RISE SICS, Report, Slides
   
6. Stream Processing in the big data era, Ahmad Al-Shistawy, RISE SICS, Report, Slides
   
7. Maps for localisation, Lars Rasmusson and Erik Ylipää, RISE SICS, Slides
   
8. End of queue detection, Al-Shishtawy  A, Ian Marsh, Bjorn Bjurling, in revision,  Draft
   
9. 12 years of road traffic data, European Conference on Data Analysis 2018,  Link to conference site
   
10. Short-Term Traffic Prediction Using Long Short-Term Memory Neural Networks, Zainab Abbas, Ahmad Al-Shishtawy, Sarunas Girdzijauskas, Vladimir Vlassov, IEEE International Congress on Big Data 2018,  Paper
    
11. Towards unifying stream processing over central and near-the-edge data centers, Sajjad HP, Danniswara K., Al-Shishtawy A., Vlassov V. In Edge Computing (SEC), IEEE/ACM Symposium,  Oct 2016 (pp. 168–178), Paper

Recorded videos

1. Congestion build up around Stockholm over 12 hours QueueBuildup
2. 16 minute scrolling plot of speed and flow data along the E4N in Stockholm Plot

Visualisation

1. Javascript visualisation of congestion

Presentations

1. Vinnova overview presentation final workshop
2. Traffic data analysis using neural networks
3. HopsWorks

Code

1. Flow_analysis (ipython notebook, save as ipynb)

Masters theses

1. Thorsteinn Thorri Sigurdsson. Road traffic congestion detection and tracking with Spark Streaming analytics, KTH, 2018, Thesis, Presentation.
2. Jón Reginbald Ivarsson,, Scalable System-Wide Traffic Flow Predictions Using Graph Partitioning and Recurrent Neural Networks, Thesis.
3. Cosar Ghandeharioon, An evaluation of deep neural network approaches for traffic speed prediction, Thesis.
4. Nathan Rengifo. Detection and Classification of Anomalies in Road Traffic using Spark Streaming, Thesis.