Finally, the event called Datathon happened at CIRM, Marseille. The Datathon was an experimental event in which the ESRs should present and work with datasets from different disciplines. For some of us (Julia, Mel and Vinicius), it was the first i-CONN event in person. It was a great experience.
Figure 1 – i-CONN members that attended the Datathon. We went to a hike in the Calanques National Park
Figure 2- First lecture by Demian Battaglia: A datathon – what is it and why do we need it?
On the first day, some ESRs presented their datasets. Mel and Julia presented theirs about sediment cover in the River Carron catchment and pharmaceuticals in rivers in the UK, respectively. Chris presented his dataset on brain networks. We watched other presentations about datasets on ecogeomorphologic data, interpersonal relationships in teamwork, global deforestation, public-private partnership contracts in Guatelama, gender bias etc. After being aware of all datasets that we could play with, we were tasked with forming a group to work with along the week.
Figure 3 – Christodoulos was describing his dataset on brain networks and Mel and Julia were considering how this can be related to river networks
Julia’s and Mel’s datasets were similar: river networks transporting sediments/chemicals. The ‘river people’ (John, Julia and Mel) watched the other presentations and thought that Chris’ dataset about the brain was somehow interesting and related to rivers because brains also have flows (in this case, flows of information) and a network structure in which space and distances are relevant. The methods used by Neuroscience to analyse brain networks are more advanced than the methods used by Geosciences to analyse river networks. Therefore, the ‘river people’ (geoscientists) quickly decided to contact the ‘brain people’ (neuroscientists) and we created a group with brain and river experts: Christodoulos, Vinicius, Julia, John and Mel. We called ourselves ‘The Brain Stream Buoys group’.
Figure 4 – Discussing similarities between river and brain networks
We talked about similarities and differences between river and brain networks. We realised that all of us were interested in analysing signal propagation through a physical structure (a network).
Vinicius suggested creating an automata model to simulate the transport of sediments/information and Chris suggested Information Theory (Mutual information and Entropy) to analyse signal propagation. We decided that we should start analysing tree-like networks as rivers and later we could modify the network structure to evaluate how the signal propagation is affected in different structures.
Figures 5 and 6 – Brainstorm of generic cellular automata model and application of information theory for brain and river networks
We spent two intense days creating the model, learning/explaining how it works to people from different backgrounds, creating options to simulate river and brain networks, and testing for some different network structures. We finally had a real collaborative and interdisciplinary work in the i-CONN. The next step would be to apply the model for our own datasets as we didn’t have time to do this in two days.
Figure 7 – Working with people from different backgrounds requires to explain several times concepts that might be simple for some and complex for others
The learnings from the event were great, especially for the geoscientists, since it was easy to discuss and learn how to apply methods from neuroscience to geosciences. Also, our colleagues from neuroscience had better programming skills and inspired the rest of us.
In summary, being forced to create a project among different disciplines was interesting and could bring great ideas. We think it would be great to have another Datathon and change groups to mix other disciplines and see different outcomes.
Figure 8 – i-CONN members appreciating the view of the Calanques
Figure 9 – i-CONN members leaving CIRM after a great week and hoping to have another Datathon soon!
By Christodoulos, John, Julia, Mel and Vinicius
