Training Workshop - GIS and Advanced Statistical Methods

Dates: 4th - 8th January 2016
Location: ULB, Brussels, Belgium
Local organisers: Goulven Laruelle, Ronny Lauerwald, Pierre Regnier, Emily Mainetti-Cloarec


 TW1 group photo: all participants and lecturers, Pierre Regnier (coordinator), Emily Mainetti (PM) and Leo Rodrigues (TM)

Summary of the workshop:

From the 4th to the 8h of January 2016 the first C-CASCADES workshop on GIS and advanced statistical methods, was held at l’Université libre de Bruxelles (ULB) in Belgium. All 15 C-CASCADES Early Stage Researchers (ESRs) were in attendance, along with 7 external participants. The workshop worked well in demonstrating how these techniques can be applied to solving actual problems in the field of Earth System Science.
The theme of the first two days was the use of Matlab to perform univariate and multivariate statistical analyses such as correlation and multiple regression analysis. Dr Goulven Laruelle (ULB), the session leader, catered well for both novice and expert matlab users by giving participants the option of creating their own scripts or modifying pre-prepared scripts.
The next theme was Time-series analysis led by Dr Sandra Arndt from the University of Bristol in which we learned how best to filter data in order to remove noise and hone in on the most important source of variation.
Just as our brains were reaching full capacity, Dr Peter Landschützer (ETH Zürich) then led a session on neural networks in which our own neurons tried their admirable best to conceptualise a Self-organizing map (SOM). What worked particularly well was a comparison of the use of neural networks and multiple regression to analyse the same dataset.
The last part of the workshop was dedicated to Geographical Information Systems (GIS) and was led by Dr. Ronny Lauerwald (Université libre de Bruxelles). This session started with a presentation explaining the basic properties and usages of GIS. We then got to produce a map of watersheds in Puerto Rico working with digital elevation models and using various tools such as zonal statistics, raster to vector and reclassify. Dr. Lauerwald ensured everyone followed and succeeded with each step of the process. Participants with some previous experience in GIS were also happy to lend a helping hand to anyone in need of some guidance.
This was the first time all C-CASCADES ESRs met and despite our diverse backgrounds and personalities we all connected exceptionally well. We also connected well with the external participants and the atmosphere in the group was great throughout the entire workshop. After the working day, we explored Brussels’ nightlife, enjoyed tasty meals as well as tried some famous Belgium chocolate and beers. At the end of the week much work had been performed, many new friendships had been made and countless of laughs had been heard. When leaving Brussels we were all tired, but already excited about our next meeting.

written by Adam Hastie and Anna Nydahl

List of the participants:

1. Anna Canning (ESR1 - C-CASCADES) - KM Contros
2. Åsa Horgby (ESR2 - C-CASCADES) - EPFL
3. Jo Snöälv (ESR3 - C-CASCADES) - UNEXE
4. Anna Nydahl (ESR4 - C-CASCADES) - UU
5. Audrey Marescaux (ESR5 - C-CASCADES) - CNRS-IPSL
6. Marie-Sophie Maier (ESR6 - C-CASCADES) - ETHZ
7. Andreas Androulakakis (ESR7 - C-CASCADES) - Deltares
8. Timo Niepel (ESR8 - C-CASCADES) - ETHZ
9. Jens Terhaar (ESR9 - C-CASCADES) - CNRS-IPSL
10. Matteo Puglini (ESR10 - C-CASCADES) - MPG
11. Adam Hastie (ESR11 - C-CASCADES) - ULB
12. Philip Pika (ESR12 - C-CASCADES) - UNIVBRIS
13. Andre Nakhvali (ESR13 - C-CASCADES) - UNEXE
14. Simon Bowring (ESR14 - C-CASCADES) - CNRS-IPSL
15. Fabrice Lacroix (ESR15 - C-CASCADES) - MPG
16. Alizée Roobaert - ULB
17. Anastasiia Ignatova - ETHZ
18. Atta Ajayebi - University of Exeter
19. Hailong Zhang - ULB
20. Katharine Maussen - ULB
21. Mehmet Kutluay - VU University Amsterdam
22. Sophie Berger - ULB

List of lecturers:

- Dr. Goulven Laruelle (ULB): Basics and Matlab, Multivariate Statistics
- Dr. Sandra Arndt (UNIVBRIS): Time-Series Analysis
- Dr. Peter Landschützer (ETHZ): Neural Networks
- Dr. Ronny Lauerwald (ULB): GIS