Synthetic data is a solution to usher in development of increasingly personalised wellbeing and health services. In the public sector, synthetic data can be leveraged to develop more comprehensive schemes for client guidance as well as serving as a tool in the service design process. AI-generated data can also make it possible to test out innovations in services and products in a much more flexible fashion than before. This calls for multidisciplinary cooperation between businesses and research establishments.
Synthetic data looks and behaves like real data, but can never expose the personal details of individuals.
Synthetic data can be generated from real datasets, using a variety of mathematical models or machine learning techniques to mimic real-life data. Synthetic data may be employed to protect the personal data of individuals by using it as “substitute data,” or alternatively to augment underrepresented groups by generating additional data points. Therefore, making good use of AI-generated data can create new business opportunities.
Pictured from left: Taija Lappeteläinen, Joonas Tuomikoski, Timo Kuisma, Rami Luisto, Toni Ruohonen and Kaisu Martinmäki
The Wellbeing DataLab project hosted an event and workshop for businesses on 3 October. It demonstrated that synthetic AI-generated data is of interest to businesses active as digital service providers and in the data sector, including those working with clients in the exercise and wellbeing fields. Synthetic data is necessary to resolve the problems inherent in collecting, sharing, and combining real-world data.
The panel discussion at the event was inspired by current challenges pertaining to using real-world data.
-Polar’s heart rate monitors and other sensors are constantly collecting a vast quantity of data from our customers. It is good there are data protection laws and regulations to ensure the rights of citizens, and we actively comply with them in our operations. Some of our customers do not want their data to be utilised in product development, said Kaisu Martinmäki, a senior researcher with Polar Electro.
-Furthermore, the regulations require that a user must be allowed to access all of their personal data in a machine-readable form. This has meant significant additional effort for us, because the data is being accumulated over many years, at times even every second.
Toni Ruohonen, head of research, development, and innovation with the Wellbeing Services County of Central Finland, said the social services and health care field are in need of quickly applicable, evidence-based research and development, in areas including the more comprehensive management of client chains.
-It can take a year, easily, to work through the various stages of authorisations and data collection before we can start to make use of data.
Rami Luisto, in charge of health technology AI development at Digital Workforce Services Ltd, added that combining data is not only a difficult technical problem, but one that is compounded with additional data sources.
-We might get partial data from three different systems, each with a slightly different administration style and regulations.
Collating the final, useable data set requires time-consuming investigative work, and an idea that emerged at launch may be five years out of date before the data has been collected and is ready for processing.
Espoo-based Coach4Pro Ltd is working on a remote management platform used by clients to provide occupational welfare services, lifestyle guidance, and exercise coaching. Timo Kuisma, manager of Coach4Pro’s public sector accounts, said that with them, service providers can make use of data to bring about immediate improvements in the services provided to customers. The operative data sample is not necessarily statistically representative, but neither does it need to be.
-We can fix problems immediately when we detect them. In my opinion, there is room for development in measuring genuine customer satisfaction and customising services based on customer feedback.
There is demand for Finnish data generation and validation
The business event’s panel discussion highlighted the fact that synthetic data will speed up the processes and development work done both within business and research. To grow businesses and enable new activity, businesses and research institutions must work together.
According to Kaisu Martinmäki, synthetic data could be used to test the usability of different data pipelines, i.e., how the source data is read, altered, and analysed, and how it gets refined into data products.
Rami Luisto said synthetic data does not need to be entirely compliant from the start. Rather, in the pilot phase it is often enough to have data that is “somewhere in the ballpark.”
-We must keep in mind there is some noise among the data, but it nevertheless lets us get to testing out various things quickly.
Toni Ruohonen said the Wellbeing Services County of Central Finland wants to develop the culture towards quicker prototyping and tests and moving things forward more rapidly.
-A wellbeing services county is not exactly awash with resources for research, which is why we need to collaborate with a variety of businesses and actors.
He emphasised that in Finland, there is a need for domestic actors to work on data generation and validation in a secure environment.
Luisto, too, hoped to see more Finns working on NLP, computing-based language AI.
At the event, companies were sparred about the possibilities of using synthetic data.
Getting right to work with pilot projects
The purpose of the Wellbeing DataLab project is to support the development of an ecosystem in Jyväskylä through the promotion of attractive conditions for developing new products. The business event saw a presentation of the Data Incubator created by the University of Jyväskylä Faculty of Sport and Health Sciences and the Faculty of Information Technology. It constitutes an entirely new experimental environment for processing exercise and wellbeing data.
To generate synthetic data, the project has collected a variety of sample data from a group of individuals training in long-distance running, as well as a control group, over one year of monitoring. The test subjects’ training was monitored using e.g. Polar’s wearable technologies. Exercise and wellbeing data was collected both under laboratory conditions at the Faculty of Sport and Health Sciences as well as in the course of the test subjects’ daily lives.
In addition, the development of the synthetic data generation tool has made use of the pilot dataset collected earlier by the Finnish Institute of High Performance Sport KIHU.
The objective of the Wellbeing DataLab project is to see the synthetic data generation tool adopted and used by businesses in the field. The project also seeks to stimulate conversations about the utility of synthetic data, and to connect actors in the business and research fields in order to engender wider cooperation in the future.
Project Specialist Taija Lappeteläinen encouraged the participants in the event to liaison at will with the project’s specialist staff, or directly with the University, if they are interested in carrying out a pilot study or tests using the Data Incubator’s test platform.
The Synthetic Data Business Event on 3 October featured, in addition to Kaisu Martinmäki, Rami Luisto, Timo Kuisma, and Toni Ruohonen, speeches from researcher Joonas Tuomikoski, project manager Maria Sukanen, and exercise expert Vesa Laatikainen-Raussi, as well as Neuwo CEO Johannes Harju, who described how Onion Sport, the Platform for Mental Wellbeing and Performance for Young Athletes, has made use of Neuwo’s technology.
The Wellbeing DataLab – Synthetic Wellbeing Data Incubator project is a joint effort by the University of Jyväskylä Faculty of Sport and Health Sciences and Faculty of Information Technology, as well as the City of Jyväskylä. Collaborating partners in the project include Polar Electro Ltd, Tampere University, and Coach4Pro Ltd. The project works to develop synthetic data derived from research results related to exercise and wellbeing, and its purpose is to support the development of future innovative products by businesses in the field. The project is partially funded by the European Union (ERDF) and the funding has been granted by the Regional Council of Central Finland.
