Learn how we tackled the tech challenges involved in processing a large volume of noisy, geospatial data, using FOSS and cloud computing. Large volumes of geotagged network data, are generated everyday using sensor devices. If processed, these data could have an important impact in location assessment and network design and planning. The main goal of this project was to identify Areas of Coverage (AoC) of telco antennas, within a large volume of noisy, point data. This brought two technological challenges: to find an algorithm that enabled us to detect the AoC, discarding the points which are not relevant; and to run this algorithm at scale, processing a large dataset, which covers the entire UK. The two challenges were tackled using a stack of FOSS. The resulting application was virtualised into a set of docker containers, and deployed on the AWS cloud, where the processing took place. In this presentation, we would like to present the final pipeline that successfully transformed the signals in AoC, and share some lessons learned during the design, implementation and deployment process. |