Embedded Files
DSSGx Munich
In the summer of 2024, DSSGx Munich supported two exciting projects with the UN International Organization for Migration's Global Data Institute, and the National Park administration Bayerischer Wald - learn more below!
In the summer of 2024, DSSGx Munich supported two exciting projects with the UN International Organization for Migration's Global Data Institute, and the National Park administration Bayerischer Wald - learn more below!
... Want to hear from the project teams in person?
Then check-out the recording of our presentation at the Civic Data Lab (Sep 13, 2024)!
... Want to hear from the project teams in person?
Then check-out the recording of our presentation at the Civic Data Lab (Sep 13, 2024)!
Knowing where to help
Enhancing Strategic Foresight and Preparedness by estimating the impact of events on migration flows with machine learning
Knowing where to help
Enhancing Strategic Foresight and Preparedness by estimating the impact of events on migration flows with machine learning
Project background and goal
Project background and goal
To effectively fulfill its three strategic objectives—saving lives and protecting people on the move, driving solutions to displacement, and creating pathways for regular migration—the International Organization for Migration (IOM) continually seeks to optimize resources and decision-making. Strategic Foresight is essential for anticipating migrants’ needs and determining necessary actions.
Recent academic research has developed machine learning (ML) models to estimate migration flows toward Europe using openly available data sources . IOM aims to adapt these models to further the understanding of short-term migration drivers in specific countries, enhancing its Strategic Foresight efforts and improving preparedness and anticipatory actions.
IOM is undertaking a pilot project to build and deploy this model. The IOM team has completed the preliminary data preparation. The DSSGx Munich Fellows will assist with the next steps: improving data quality, training ML models, evaluating them, and advising IOM on the next steps thereafter.
Our impact
Our impact
Anticipating the impact of events on migration drivers and cross-border migration flows is crucial for assessing the changes and challenges IOM operations may face in the coming months. Enhanced foresight on migration drivers and movements in specific countries or regions is vital for saving lives and shaping policy solutions for regular migration pathways. This, in turn, improves the data-driven allocation of IOM’s resources. Through better resource allocation, IOM and partner UN organizations can provide better services to migrants and internally displaced people.
Partner organization
Partner organization
The team
The team
Masabah Bint E Islam
María Belén Arvili
Derya Durmush
Jorge Roa
Valerie Gastner
(Project partner)
Results
Results
The project team gathered extensive migration data, incorporating diverse factors such as socioeconomic conditions, environmental variables, and demographic trends. By using machine learning models and predictive analytics, the team developed a sophisticated tool to visualize and forecast migration patterns effectively. The solution was built with modularity, allowing IOM to adjust parameters based on real-time data inputs, and it included dashboards that provided clear, actionable insights for IOM stakeholders. This tool enabled IOM to refine their migration response strategies, allowing for more proactive and targeted interventions tailored to specific migrant demographics and regions at risk.
Learn more about the project in our final presentation:
20240925_DSSGx_Munich_2024_Closing_IOM_Slides_approved.pdfHarmonizing tourism and nature protection
Data-driven visitor management in the National Park Bavarian Forest
Harmonizing tourism and nature protection
Data-driven visitor management in the National Park Bavarian Forest
Project background and goal
Project background and goal
The Bavarian Forest National Park is the oldest national park in Germany and a popular tourist destination for recreational activities, with more than 1.4 million visitors per year. Its popularity, however, also presents logistical challenges. Forecasting, managing, and controlling tourist flows is imperative to ensure sustainable tourism and the use of natural resources.
While much data about tourist flows is collected at and around the Bavarian Forest, it is quite heterogeneous and stored in different formats. Further automation of the processing and integration of the different data sources is also needed to enable a seamless information flow.
The goal of the project is to support different stakeholders with data integration and dissemination of information about the tourist flows in the Bavarian Forest National Park.
Our impact
Our impact
The project's success will enable a better information flow to different stakeholders and thereby help ensure sustainable resource management, including reducing pollution and CO2 emissions in and around the Bavarian forest.
Partner organization
Partner organization
The team
The team
Ayesha Younas
Manpa Barman
Anthony Garove
Patricio Ferreira
Max Mangold
(Project partner)
REsults
REsults
The project created a data-driven visitor management solution for the Bavarian Forest National Park, aiming to balance high visitor traffic with ecological preservation. The team built an end-to-end pipeline integrating diverse data sources, such as visitor counts, weather, and parking data, while addressing data quality and warehousing challenges. Using an Extra Trees Regressor, they developed a predictive model to forecast visitor flows, covering the entire park and specific sub-regions. This solution included a user-friendly dashboard, providing real-time insights to multiple user personas—visitors, employees, and the park’s management team. The dashboard helped optimize resource allocation, anticipate peak days, and improve visitor experiences by reducing crowding. The outcome supports sustainable tourism management, minimizing human impact on the ecosystem, and serves as a replicable model for protected areas with similar needs
Page updated
Google Sites
Report abuse