Improving safety management through data mining – AI Safety
- The development of occupational safety and well-being at workplaces requires accurate information on the most effective methods and measures. This study analyses large data sets in order to produce new information (weak signals) and use it in order to obtain better results from development activities.
- Current data processing methods can already predict, for example, the number and severity of occupational accidents, but weak signals predicting danger are still poorly understood in safety research. The integration of different types of data held by organizations with occupational safety data and the study of this combined data using data mining and machine learning methods can produce new kinds of information for safety management needs.
- The study will use machine learning methods to find new information on how organizations should develop data collection related to their operations from the perspective of occupational safety. This is necessary in order to make the best possible use of data mining and machine learning in safety management, for example, in order to improve the focusing of safety measures. For this purpose, the strengths and weaknesses of current data, knowledge gaps, silent signals that may be found in data and the connection of different data to each other from the perspective of occupational safety are examined.
- The AI Safety research project specifically aims to develop the application possibilities of machine learning and data mining in safety management in order to facilitate the identification of work-related hazards and methods to combat them.
Materials and techniques
Four Finnish industrial companies from different sectors are involved in the study, and their occupational safety and occupational safety management are examined in a data-driven manner. In addition to data analysis, interviews and workshops are carried out for different personnel groups of the participating companies.
Use of employee information in research
Brief for employees
The AI Safety research project of the Finnish Institute of Occupational Health involves four Finnish industrial organizations specified on this website.
The purpose of the project is to develop safety management and occupational safety. The study uses and combines the data collected and accumulated by the participating organizations from their operations at employee and unit levels.
The organizations provide the research project with data. These data contain identifiable information on the current and former employees of the organizations and, possibly, employees of other employers active in their respective areas.
All types of personal identifiers, such as names, will be removed from the data before analysis. The data used in the study are, depending on the organization, from 2010–2022.
Employers are not provided with combined or analysed data that can be used to identify individual people. All research results will be reported in a way that makes it impossible to identify individual people.
The employees of participating organizations have been provided with a detailed brief of the study, the use of the employees’ identifiable information and the employees’ rights concerning the use of their identifiable information by e-mail.
Company-specific briefs can also be obtained from the following links:
- Prefere Resins Finland Oy
- Sappi Finland Operations Oy
- Metso Outotec Finland Oy
- Konecranes Finland Oy
If you have any questions about the study or how your information is used in the study, please contact the researchers of the project:
Tarja Heikkilä, specialist,
030 474 8636
Maria Tiikkaja, research manager,
030 474 2750
Project manager, research manager
Maria Tiikkaja, Eero Lantto, Vuokko Puro, Tuula Räsänen, Tarja Heikkilä, Simo Virtanen, Jukka Kärkimaa, Kimmo Sirén
The project is funded by The Finnish Work Environment Fund, Finnish Institute of Occupational Health and co-operating organizations: Sappi Finland I Oy, Metso Outotec Finland Oy, Konecranes Finland Oy and Prefere Resins Finland Oy.