Between 19th and 21st February 2019, Munich played host to the Big Data from Space 2019 (#BiDS19) conference. The conference was jointly organised by ESA, SatCen and the JRC and hosted by DLR (the German Space Agency). It’s remit was to cover all spatio-temporal Earth and space observation data collected by a variety of sensors, and to investigate all aspects of data management, processing and exploitation.
This was the first time that I had attended this biennial conference, and I was really impressed. There were some top notch presentations which came from a mix of academics, commercial entities, and governmental and governance organisations and included technology providers, data providers, start-ups, industry leaders and more. Overall it felt a very inclusive community of people interested in all aspects of geospatial and space science.
And the coffee and cakes were excellent (priorities, people!!). However, one thing that came to my notice was that the coffee seemed much stronger in the afternoon break – what I never figured out was whether this was a purposeful change or whether I was so loaded on coffee and sugar by that time that it only seemed stronger????? Mysteries!
In the blog post I just want to quickly list some of my highlight presentations (there were also very good posters and demos but I won’t be going into those in this post).
Day 1 got off to a brilliant start with a fascinating keynote from Xiaoxiang Zhu, the Head of EO Data Science at DLR and the Technical University of Munich which was titled “Artificial Intelligence and Data Science in Earth Observation (EO)”. One of the things that struck a chord was the use of generative adversarial networks for cloud removal in Sentinel-2 imagery using Sentinel-1 data. The second keynote speaker, Volker Markl from the Technical University of Berlin Data Analytics Lab built on the first keynote by discussing “Mosaics in Big Data: Stratosphere, Apache Flink, and Beyond”. This was the first time I had come across Flink, and it was interesting to see how it performs in comparison to other members of the Apache family. Marcus Netler gave a very interesting talk about Actinia a system being developed for cloud based geoprocessing. Unfortunately I had to duck out of the day halfway through this talk which meant I also missed two (reportedly) very interesting talks on Jupyter and Pangeo: I had a date with the Bayerische Staatsoper (first world problems!)*.
Day 2 started by catching me unawares: I was not expecting to be so engaged by a presentation about Quoting and Billing but Ingo Simonis from the OGC had my attention from the first slide and never lost it. Some of the work undertaken by the OGC on behalf of the geospatial sector is truly inspiring and incredibly important. The entire session on Analysis Ready Data (ARD) was immensely relevant to everyone there. There were talks about the Austrian and Australian data cubes and how they are being used to drive new analytics, as well as a perfectly timed and delivered presentation from Peter Baumann about how we need Analysis-Ready Services rather than ARD.
It was another corker to start Day 3: the keynote from Gustau Camps-Valls (Universitat de València) about “Machine learning in Earth Observation data analysis” blew my mind as he proceeded to list the wide variety of state-of-the-art advanced machine learning techniques that can be applied to EO data. That one person can be involved in so many areas was truly inspiring (seriously, that website I link to doesn’t do him justice. I think most people in the session were on the edges of their seats when he was presenting). Philipp Hochreuther gave another fascinating presentation on the bottlenecks involved in big EO data processing, which is something we all probably need to be aware of but rarely think about.
Thankfully there was only one joke about Brexit (and it was quite a good joke!) but it annoys and saddens me to think that this may be my last trip to continental Europe as a European citizen. There is so much inspiring (and awe inspiring) work being done in this field that I seriously hope that we scientists (whether academic, commercial or governmental) can continue to work together as part of such a wonderful community (despite what daft things the UK might be doing).
This was a great conference and I thank all the speakers and the organisers. I look forward to 2021 in Romania!
*BTW, Nabucco was excellent.