Understanding accessibility to basic services for public housing in Milan: An open source approach

This is the repository for the Undergraduate Thesis of Giulio Frey.

Aim of the project

We want to analyze how connected are different public housing with essential services around the city of Milan.

Abstract

We propose a connectivity index built using fully open-source tools to evaluate the accessibility to basic services of the public housing stock owned by Comune di Milano. Our approach is scalable to other cities, fully reproducible and has negligible costs of implementation. We cover the inner workings of the routing machine necessary to build the index, qualitatively describe the housing stock and go over a comprehensive history of public housing institutions. Results of the connectivity index are analyzed using the Gini index, correlation and mutual information with distance from the city center. We find that most services are quite equally distributed among the public housing units when looking at the Gini index. All services are negatively correlated with distance from the city center and university buildings and cultural points of interest have the highest values of Mutual information.

Contents

code

This folder contains all the code that is used in the project. The main task is the creation of the distance matrix using OSRM and subsequent analysis.

• analysis.ipynb: performs the analysis
• data_cleaning.ipynb: creates the datasets for essential services around the city.
• calculator.py: constructs the distance matrix using OSRM
• calculator_fixer.py: constructs the distance matrix using OSRM but only of one specific dataset. It was necessary as we faced geoencoding issues with one dataset.
• merger.ipynb: merges the fixed dataset with all the others.
• mm_geofactor.ipynb: creates the social housing dataset. Converts the pdf provided by MM in a .csv file and the geoencodes the addresses
• tt_matrix.ipynb: experiments in calculating the time distance matrix between social housing and essential services using OSRM
• visual.ipynb: creates visual to support the analysis

outputs

This folder contains all the datasets created for the project

• calc_dataset.csv
• calc_fixed_df.csv
• fixed_dataset.csv
• mm_dataset.csv
• services_df.csv
• calc_dataset_backup.csv
• distr_df.csv
• index_df.csv
• row_file

Data sources

We use datasets provided by the portal Open Data of the Milan municipality in order to track essential services.

A record of Social Housing in the city of Milan provides physical addresses. It can be found here