Design and Implementation of a Scalable Crowdsensing Platform for Geospatial Data

In the recent years smart devices and small low-powered sensors are becoming ubiquitous and nowadays everything is connected altogether, which is a promising foundation
for crowdsensing of data related to various environmental and societal phenomena. Very often, such data is especially meaningful when related to time and location, which is
possible by already equipped GPS capabilities of modern smart devices. However, in order to gain knowledge from high-volume crowd-sensed data, it has to be collected
and stored in a central platform, where it can be processed and transformed for various use cases. Conventional approaches built around classical relational databases and
monolithic backends, that load and process the geospatial data on a per-request basis are not suitable for supporting the data requests of a large crowd willing to visualize
phenomena. The possibly millions of data points introduce challenges for calculation, data-transfer and visualization on smartphones with limited graphics performance. We have created an architectural design, which combines a cloud-native approach with Big Data concepts used in the Internet of Things. The architectural design can be used as a generic foundation to implement a scalable backend for a platform, that covers aspects important for crowdsensing, such as social- and incentive features, as well as a sophisticated stream processing concept to calculate incoming measurement data and store pre-aggregated results. The calculation is based on a global grid system to index geospatial data for efficient aggregation and building a hierarchical geospatial
relationship of averaged values, that can be directly used to rapidly and efficiently provide data on requests for visualization. We introduce the Noisemap project as an exemplary use case of such a platform and elaborate on certain requirements and challenges also related to frontend implementations. The goal of the project is to collect crowd-sensed noise measurements via smartphones and provide users information and a visualization of noise levels in their environment, which requires storing and processing in a central platform. A prototypic implementation for the measurement context of the Noisemap project is showing that the architectural design is indeed feasible to realize.