roberto-nai/ANAC-OD-ETHICAL

Machine learning in procurement with a view to equity

Machine learning in procurement with a view to equity

> Directories

config

Configuration directory with config.yml.

maps_data

Directory with Italian region heatmaps.

ml_plots

Directory with ML results plots.

ml_results

Directory with ML results.

open_data_anac

Open Data to be copied in this directory for starting the experiment are available here: https://bit.ly/4dkNMeQ.

stats

Directory with procurements stats.

utility_manager

Directory with utilities functions.

> Script Execution

01_data_filter.py

Loads the various datasets (in CSV format) and generates the event log. Only keeps cases starting with the TENDER_NOTICE event.

02_data_filter.py

Loads the filtered dataset in the previous script and trains Machine Learning (ML) models.

03_data_plot.ipynb

Loads the ML results ad plot the results.

04_data_map.ipynb

Loads the the various datasets (in CSV format) plot values by Italian region on a map (use limits_IT_regions.geojson).

> Configurations

conf_cols_filter.json

List of columns (features) to be filtered.

conf_cols_stats.json

List of columns (features) to be filtered.

conf_cols_type.json

List of columns (features) types.

limits_IT_regions.geojson

GeoJSON file of Italian regions (used by 04_data_map.ipynb).

> Script Dependencies

See requirements.txt for the required libraries (pip install -r requirements.txt).