Predicting Asylum Application Outcomes in the European Union

Due to many ongoing conflicts around the world today, forced international migration continues to be a vital event in many people's lives. In this project, I hope to develop effective model that can predict future asylum applications and help us better understand what factors affect their eventual outcome. This project uses Eurostat's migr_asydcfstq dataset with quarterly asylum application outcomes to the EU¹. As of time of publication, the information from Q3 2023 is still partial, so it has been excluded from study.

This project is built in Python 3.11.4 and its respository is organized into the following sections:

1. figures/ -> All figures for data visualization
2. results/ -> Trained models
   • Some explanation of the significance of each can be found in model_notes.txt
3. report/ -> Reports and presentions on methodology and the pipeline
4. src/ -> Contains all the code for:
   • Exploratory Data Analysis in eda.ipynb
   • Data Splitting, Preprocessing, and Model Development in model-development.ipynb
   • Model Evaluation and Interpreting in model-evaluation.ipynb

The packages used in this project are listed below. You can download all these dependencies into a new conda environment by running conda env -n migr create -f environment.yml and then conda activate migr.

• pandas 1.5.3
• numpy 1.24.3
• matplotlib 3.7.1
• seaborn 0.12.2
• scikit-learn 1.3.0
• shap 0.42.1
• joblib 1.3.2
• xgboost 1.7.6

_{¹This includes the 27 EU member states as well as affiliated states like the UK (with data up until 2020), Montenegro, Iceland, Liechtenstein, Norway, and Switzerland}