I'm Monserrat Lopez! I am a Mexican economist currently pursuing a Master’s in Data Science for Public Policy at the Hertie School. I have over six years of professional experience conducting data-driven research and policy analysis. My passion is to use data for social good and to find policy solutions. In this repository, I want to save and share my Data Science learning journey and my work.
-
Description: This project was the final work for a Machine Learning course at the Hertie School. With the help of my team, we developed and tested three different Machine Learning Classification Models to predict discrimination faced by individuals seeking jobs after resettlement in Europe.
Tools/Technologies: Python (Pandas, NumPy, Matplotlib, Scikit-learn).
Results: For predicting migrant discrimination while looking for a job in Europe, we evaluated three different Machine Learning Classification models: Logistic Regression, Random Forest, and XGBoost. The Random Forest Classifier was the best performer, with higher accuracy and superior AUC value, achieving an accuracy of nearly 73% after hyperparameter tuning. The nature of the data, the incompleteness of the information, and the intrinsic challenge of predicting the complex phenomenon of perceived discrimination can affect this percentage. However, we managed to reach a model with a sufficiently high level of accuracy to make reliable predictions on European job market discrimination towards minority groups.
-
Description: This app was developed as a final team project for the Data Structures and Algorithms course at the Hertie School. We created an application that assists users in managing and settling their debts effectively. I worked mainly in the front end design of the application.
Tools/Technologies: Python, Flask, HTML & CSS.
Results: You can visit our final prototype here: https://elena3er.pythonanywhere.com/
-
Description: This project is part of my DataQuest Portfolio and aims to use a Machine Learning model (Ridge Regression) to predict local weather, using weather data from a station at the Oakland International Airport.
Tools/Technologies: Python (Pandas & Scikit-learn).
Results: In this project, I utilized Ridge regression models for predicting maximum temperature based on historical weather data. The Mean Absolute Error (MAE) of the first estimated model was approximately 3.4 degrees Fahrenheit. However, by including more predictors in the model, the MAE decreased to approximately 3.3 degrees Fahrenheit.