Skip to content
Public sector AI Data analytics Cloud solutions

Case Fintraffic

Artificial intelligence improves the quality of traffic volume data and enhances the work of experts 

Digia's business intelligence unit, Productivity Leap, built a traffic data reporting system for Fintraffic. A machine learning-based correction application improves the quality of the data. The new reporting system makes operations more efficient, and high-quality traffic volume data benefits society. With the new reporting system, information will be openly available online for everyone who needs it. 

 

Fintraffic is a special assignment group operating under the ownership and steering of the Ministry of Transport and Communications. It offers and develops traffic control and management services in all modes of transport and responsibly ensures the safety and smooth flow of traffic. One part of the service produced by Fintraffic is the collection of traffic volume data, which is implemented with the help of automatic traffic measurement points (LAM). There are about 500 stations in Finland. 

Traffic volume data is of great importance to the surrounding society. Various administrative and official bodies utilise the information in a wide range of ways, from traffic planning to political decision-making, and the media is also interested in traffic information. Due to the versatility of data utilisation, it is important that the information is of high quality and accurate. 

"LAM data and the reporting system play a key role in the service we provide to the Finnish Transport Infrastructure Agency and, ultimately, to citizens. The reliability of data and low-threshold access to it are very important to us," says Eetu Karhunen, Project Manager and Service Manager of the traffic measurement service. 

What we did

  • LAM measurement point reporting 
  • Data repair application 
  • Open data distribution 

What we used

  • AWS
  • Snowflake
  • Agile Data Engine
The solution is a pioneer in bringing machine learning and artificial intelligence to help human work. Now we have a system that stands the test of time and can be scaled as needed.

Eetu Karhunen, Service Manager, Fintraffic

User-friendliness and quality as a starting point

Fintraffic's needs were clear: The old system, which had reached the end of its life cycle, had to be renewed, and the corresponding functionalities had to be included in the new reporting system. At the same time, the system's usability and information quality had to be improved. Productivity Leap was chosen as the partner, and its experience and precise expertise convinced Fintraffic.

"There were people in the team who were familiar with the previous system, so it was easy to start cooperation," Karhunen says. 

The old Tiira system was replaced by a modern, cloud-based solution where reporting is handled with easy-to-use Microsoft Power BI reports. Anyone can openly use basic traffic information through the public Digitraffic.fi online service and embeds of public reports can also be found on Fintraffic's website. Experts who need more extensive and refined data have access to more detailed reports subject to permission. 

The new system's ease of use is reflected in the number of users and their feedback. 

"The new reporting system is much more user-friendly. As a result of the improvements, we have gained more users, and the user feedback has been positive," Karhunen says. 

The repair app leverages artificial intelligence

One significant improvement brought about by the new system is a correction application based on machine learning, which makes data of higher quality and more accurate and makes the work of experts working with data more efficient. The application identifies deviations in the data of individual measuring points, for example, due to a technical fault, and suggests corrections based on modelling based on the metering point's previous traffic volume history. The application utilises continuously accumulating data and refines its correction suggestions.

In the previous system, the identification and correction of data anomalies relied entirely on manual expert work. The new application significantly reduces the daily working time spent on repair work. Now, it is up to the expert to either accept or reject the application's correction suggestions. 

"Machine learning supports and enhances the work of a trained expert. Work that previously took several hours can now be done in a quarter of an hour," says Kimmo Saastamoinen, who works as an expert consultant in traffic calculations.  

"We haven't wanted to go for a fully automated solution yet, but in the future, we may be able to increase the degree of automation." 

The repair application also supports the maintenance of LAM measuring points. Technical errors that reveal faults at measurement points are detected faster in the data, which speeds up the initiation of corrective actions and contributes to improving data quality. 

"With the help of the application, we can better control our maintenance operations," says Karhunen. 

Digitalisation supports the work of experts

With the new reporting system and repair application, Fintraffic has taken a significant step in the field of digital transformation.

"The solution is a pioneer in bringing machine learning and artificial intelligence to help human work. There is a lot of interest in the topic, and we also receive a lot of internal inquiries about the solution," Karhunen says.

Karhunen is pleased with Productivity Leap's cooperation on the project.

"Now we have a system that stands the test of time and can be scaled as needed. The project was successful in every way, as we stayed on schedule and budget and achieved the set goals. The flow of information has worked, and from the project manager's point of view, everything has gone great."