For creating sports innovations we need tools to collect data on physical activity behaviour. Collecting data is time-consuming and expensive. Therefore, sharing data and using it to answer multiple research questions is desirable. Especially in larger multidisciplinary research projects, it is beneficial if a diverse data set can be collected effortlessly. For this, we created a smartphone app that can be used for scientific data collection on physical activity behavior in sport research.
The requirement for the app was that the different research partners could contribute, use the app independently and that the individual contributions could be easily combined into a larger app. We created a base app which provides a framework for easy implementation of custom modules. Each functionality in the app is represented in a module and can be created by different parties. The app uses the EventBus framework [1] to allow easy communication between independent modules including an independent GUI. In our presentation we will further explain this independance using the beacon physical activity app BAMBEA [4] .
Over the span of three years we made a sensor collection-, emotion input-, beacon scanner-, beacon handler-, goal-, reward- and message/notification-module. We combined different modules into different versions. A first version initially focused on running behaviour and used a movement- and a heart rate-module that were originally created for research [2] and different modules of partners (like gait step frequency and emotion detection). The app was extended with a GUI to test the user experience of the modules [3]. Finally, the framework was updated and combined different modules and a new GUI for a bootcamp app [4].

In conclusion we experienced the modular framework as a good solution for reuse and collaboration within multidisciplinary research projects. By allowing for the independent development of modules only intense communication between partners was needed when conflicts occurred between modules. Another advantage was that partners could develop their module in their desired programming language and later integrate their module in the app by providing a small wrapper. We consider the framework approach as a success for projects where close collaboration is desired with a multidisciplinary team. However, we would not advise this approach when it can be expected that multiple heavy performance solutions are required and resources like data collection, storage and backup cannot be shared. In that case a tight integration and continuous collaborative implementation is likely a preferred approach.
This research is made possible by the Amsterdam University of Applied Sciences research programs Urban Vitality, Amsterdam Creative Industries Network, the Amsterdam Institute of Sport Sciences (AISS) and by the Dutch National program COMMIT/.