BigO – Big Data against childhood obesity
The high prevalence of childhood obesity is an emerging health issue, with large societal consequences. The aim of BigO was to use Citizen Science to collect and analyse Big Data on children’s health related behaviour and their environment enable to identify local risk factors affecting childhood obesity rates.
A snapshot of the project
BigO was a Citizen Science 4-year EU funded program which involved 14 partners from different countries across Europe, including schools, clinics, research organisations, public health authorities and commercial enterprises. Thus, the overall aim of BigO was to collect and analyse big data on behaviour and living environments related to childhood obesity. This information could then be used to allow public health authorities to plan and execute effective programs to reduce the prevalence of childhood obesity.
BigO collected and analysed anonymous data on children’s behavioural patterns and their living environment. By using advanced analytics and sophisticated visualisations, BigO extracts data-driven evidence on which local factors are involved, and how these factors influence childhood obesity in Europe. As an open platform, BigO was envisioned as a tool for local public health authorities. This is done in various steps. School aged children become citizen scientists by collecting data about their behavioural patterns and local environment, using the myBigOapp. This data is anonymised and used to create complex statistical models to analyse how behaviour and the environment influence the prevalence of obesity. This anonymous information can be used to predict how policy changes could influence obesity rates and can be used to compare different communities on a group level.
Specifically, BigO developed an integrated platform for measuring real-time obesogenic behaviours of children and adolescents, using their mobile devices and offers aggregated localized information from all participants to public health scientists and policy makers. MyBigO app is a tool for children’s education supporting citizen science activities. GPS and accelerometry are automatically measured while pictures of meals and of food advertisements are sent by the children. More than 30 indicators on physical activity, visits to food shops and recreational areas are calculated daily, offering objective measures for different geographies. Aggregated at geohash level, this localized behaviour information is free from identifiable elements and preserves children’s privacy. Students are recruited as citizen scientists in line with ethical guidelines and avoidance of stigmatization.
The results of the project
The BigO platform has proved robust operation with the system being up and running 24-7 in more than 30 months supporting large-scale data collection at schools and clinics. BigO reached out to more than 21000 children with 5809 using the BigO app and contributing their data. Skipping school lunches for fast food eating in Stockholm schools close to food retailers, low levels of walking in suburban areas in Thessaloniki and changes of physical activity habits during COVID 19 lock down are some of the behaviours that have been “revealed” via objective measures for the first time.
There are various types of impacts of this Citizen Science project. Technical impact which means that algorithms for extraction of useful parameters are developed and applied. Some of them are patented. The potential public health impact of such a system is shown by several cases that are presented to policy makers. Concerning the scientific impact, several academic papers are written on the technical advances, the clinical outcomes and also a vision paper and a position paper on Big Data for the study of obesity. Finally, with respect to the impact on academic education, the project is used as an example in a wide array of data-science courses and courses on food intake and childhood obesity. And last, it was used as a case during a hackathon of the European statistical bureaus.
More research: Sensory and metabolic drivers of eating behaviour