From Satellite Data to Sankt Hans: Notes from the Data for Wellbeing Workshop at Copenhagen
I (Theophilus Aidoo) recently had the pleasure of attending the Data for Wellbeing for the Most Vulnerable workshop at the ICWSM conference in beautiful Copenhagen. The workshop brought together researchers working at the intersection of social good and data science, a space that feels more important than ever.
I was there to present my work on mobility estimation using satellite imagery, that is, how we infer parking lot occupancy from 3-meter satellite imagery. But while sharing research was part of the experience, the true value of the workshop came from listening, learning, and connecting with others who care deeply about using data to improve lives.
In our project, we explore how comparing images of the same parking lot at different times can reveal subtle mobility changes. At 3-meter resolution, spotting individual vehicles is nearly impossible. But by training a pairwise comparison network to decide which image looks more "occupied," we can extract signal from all the visual noise.
The “weak supervision” part of our method comes from how we annotate the data. Instead of manual labeling, we use retail shop opening patterns in Germany as soft labels assuming that lots are likely full when stores are open and emptier when they are closed. This approach lets us scale our analysis across both time and geography. We applied it to detect changes in parking activity in Sudan before and just after the onset of war and saw a meaningful shifts in mobility that matched what we knew from the ground.
In the workshop, I listened to the wonderful works by the researchers gathered, but what truly stuck with me was the inspiring keynote by Paddy Brock, who spoke about the UN’s work in collecting real-time data for disaster response. His talk reminded me why we do this work that behind every pixel, pattern, or prediction, there are real people whose lives could be improved through better insight. It was a humbling and energizing reminder that data science, at its best, is in the service of humanity.
And of course, no trip to Copenhagen would be complete without experiencing a bit of Danish culture. My visit coincided with Sankt Hans Aften, celebrated on June 23rd, a traditional midsummer event that marks the eve of Saint John’s Day. I joined locals at Nyhavn (New Harbour), where a bonfire was lit on the water to mark the occasion. It was nice to witness the golden light stretching late into the evening.
A heartfelt thank you to the workshop organizers and my funders for making this possible. It was a privilege to share my work with such an engaged community.
You can read our paper here.
