Work on our automated firefront tracking algorithm has continued strongly over the past two weeks. Peter has been working on creating simple Kepler actors to encapsulate the process (below). The firefront tracking algorithm processes crowdsourced data aggregated by a sliding temporal window. Before implementing the firefront tracking algorithm, Peter is applying simple analytics on the tuples in the sliding window, such as calculating the sum and maximum of attributes of the tuples. Since the algorithm adopts standard spatiotemporal analysis tools, which already have open source implementations, based on this knowledge it should be possible to implement the firefront tracking algorithm efficiently.
Matt also gave a presentation at RMIT University on “Challenges for research into big geospatial data”. The presentation included a discussion of the RISER firefront tracking using crowdsourced emergency calls. The example demonstrates how crowdsourced information can provide real-time emergency infor- mation, supplementing higher quality but also higher latency authoritative information sources. A paper for submission to PNAS (Proceedings of the National Academy of Sciences) is also currently in the final stages of preparation (below).