AI summer program for social projects at LMU

This year, young data scientists from around the globe again applied to participate in the German branch of the University of Chicago’s Data Science for Social Good initiative. In the process, they helped the Bayerischer Wald (Bavarian Forest) National Park to control visitor flows more sustainably, while also helping the International Organization for Migration to make refugee movements more humane. Both projects show how artificial intelligence (AI) can play a part in resolving societal challenges.

The Bavarian Forest National Park is the oldest such park in Germany, and idyllic countryside makes it an attractive destination for more than 1.4 million guests a year. Precisely this popularity is creating headaches for the nature reserve, however, as the most heavily frequented parking lots quickly become overcrowded, especially at weekends. “When that happens, visitors simply park by the side of the road – which is prohibited – and go walking aside from the signposted paths,” says 26-year-old Manpa Barman, adding that this practice is a safety hazard and also disturbs the wildlife. In the interests of sustainable tourism, she notes, it is important to predict, manage and control the flow of tourists in advance.

IT student Barman does not work for the national park: She comes from India and is currently studying at the University of Stuttgart. Together with her Data Science for Social Good 2024 (DSSGx) team, she is using her expertise to develop a machine learning-based model to predict visitor footfall. Although 38 sensors have already been installed at the points of entry to the park, “[the data] has so far been stored in very varied ways and different formats,” she explains. However, her work has significantly improved the team’s understanding and analytics, and has thus contributed to better personnel planning and more sustainable resource management – including less pollution of the environment and fewer CO2 emissions.

Fellowships for data scientists

The DSSG fellowship DSSGx Munich is a two-month fellowship for talented data scientists, statisticians and social scientists from all over the world. The LMU branch of the program was launched in 2023 and takes place every August and September. It is funded by the Munich Center for Machine Learning (MCML) and the Bavarian Ministry of Digital Affairs. For each project, two roughly five-person data science teams spend two months each working on real, everyday problems at public and nonprofit organizations who simply lack the time and the financial resources to come up with viable solutions. The project is a local branch of the DSSG program launched by the University of Chicago, where Professor Rayid Ghani had the original idea in 2013.

Saving lives with data science

Enlarge

The second DSSGx team in Munich worked with machine learning models in collaboration with a project run by the Berlin-based International Organization for Migration (IOM), under the aegis of the United Nations (UN). Using data sources in the public domain, the aim was to predict future migration flows at the Horn of Africa, where around 30 percent of all refugee movements on the African continent took place in 2019. “Our hope is to identify the needs of migrants better and at an earlier stage, to take the steps that are necessary and to save lives,” explains María Belén Arvili. She comes from Buenos Aires in Argentina and studied issues such as urban social policies at the Universidad Nacional de San Martín.

Databases detailing past conflicts, political events and the weather laid a foundation for the work of the 33-year-old and her team. Drawing on both this and their own data, they were able to develop a model that can predict future migration flows up to six months in advance. “If migration is well managed, it can improve societies,” the humanist and computer scientist is convinced. She also believes her early-warning system can help: Her insights and recommendations are now being channeled into the ongoing work of the Global Data Institute.

Both Barman and Arvili first heard about the DSSG program from a post by the MCML on the LinkedIn network. Why were they selected? “Probably because we are interested in social issues as well as statistics and mathematics,” Arvili says. That is true. “But we also take care to bring together as diverse teams as possible from around the world,” adds program co-manager Jacob Beck. All the participants only found out which project they would be assigned to after arriving in Munich. “I was really excited about the Bavarian Forest project as soon as I heard about it. And fortunately, that is the one they assigned me to,” a smiling Barman recalls.

Different perspectives

Both participants found the work very fulfilling – not only because of the positive impact on society, but also because of the personal encounters in Munich and with the diverse teams. Mexico, India, Pakistan, the USA, Russia, Argentina: The fellows in attendance came from different parts of the world and had varying backgrounds – from sociologists to data scientists to electrical engineering specialists. “Everyone has different perspectives on the same topic,” is the consensus view. While this could be challenging at times, it was always a lot of fun and was important for the positive project outcomes.

For the duration of their two-month stay, the fellows were accommodated at student halls of residence in Munich. All agreed that the organization was excellent. In addition to shared cultural experiences, picnics in the English Garden, trips to Neuschwanstein and the Königssee lake (not to mention the obligatory visit to the Oktoberfest!), the participants were also taken to Berlin, where they presented their projects to three deputies at the German Bundestag. After a leaving celebration to conclude the project, everyone felt a little sad. The two participants are nevertheless in full agreement: “We hope that we will be able to continue working on ‘our’ projects in the future.”

Program co-leader Clara Strasser Ceballos is likewise turning her attention to what lies ahead, because “you finish one DSSGx and move straight on to the next one”, as she puts it: “So, we are already on the lookout for exciting new projects and project partners whom this initiative can help next summer.”

For more information, see:

Data Science for Social Good Munich