Exploratory data analysis on global drug phenomena — mortality, consumption, crime, and penal policy
Academic project completed as part of course 8PRO408 — Programming Tools for Data Science at the Université du Québec à Chicoutimi (UQAC), Summer 2023.
Inspired by the Netflix series Narcos, this project takes the perspective of a fictional anti-drug agency to investigate global drug phenomena through four independent studies: drug-related mortality, substance use by age group, correlation between cocaine deaths and crime rates, and drug-related incarceration trends.
Each study was driven by a concrete question an anti-drug organization might ask — validating or challenging common assumptions through data.
| # | Topic | Data Source | Notebook |
|---|---|---|---|
| 1 | Drug-related deaths by substance — Connecticut, 2012–2018 | Chief Medical Examiner's Office (Kaggle) | 01_morts_par_drogue.ipynb |
| 2 | Drug use by age group — United States | National Survey on Drug Use (Kaggle) | 02_consommation_par_age.ipynb |
| 3 | Cocaine death rate vs. homicide rate correlation (worldwide) | Our World in Data + World Bank | 03_cocaine_et_criminalite.ipynb |
| 4 | Drug trafficking & possession imprisonment trends (worldwide) | UNODC — Criminal Justice Statistics | 04_emprisonnement_trafic.ipynb |
Consolidated report → rapport.ipynb
- Fentanyl: dramatic increase in overdose deaths in Connecticut between 2012 and 2018, disproportionately affecting men
- Painkillers & benzodiazepines: leading cause of overdose deaths among women — a high-leverage intervention point through prescription monitoring and physician awareness
- Alcohol & marijuana: peak consumption among 18–23 year-olds; inhalants are predominantly used by teenagers
- Cocaine & crime: positive correlation detected in several countries (Israel, Venezuela, Russia, Spain, Canada) — no causal relationship established
- Penal policy: significant cross-country disparities in imprisonment rates for drug trafficking vs. possession, with an upward trend over 2016–2021
escobaddictions/
├── rapport.ipynb # Consolidated report
├── 01_morts_par_drogue.ipynb # Study 1 — Deaths by substance (CT, 2012–2018)
├── 02_consommation_par_age.ipynb # Study 2 — Drug use by age group (USA)
├── 03_cocaine_et_criminalite.ipynb # Study 3 — Cocaine deaths vs. crime rate
├── 04_emprisonnement_trafic.ipynb # Study 4 — Drug-related imprisonment (worldwide)
│
├── data/ # Raw datasets
│ ├── drogues/ # Substance-specific mortality rates (Our World in Data)
│ ├── drogues_sante/ # Substance use disorder statistics
│ ├── consommations_drogues/ # Behavioral drug consumption surveys
│ ├── murder/ # Homicide statistics (World Bank)
│ └── drug_related_death_classification.csv
│ drug-use-by-age.csv
│ data_cts_prisons_and_prisoners.csv
│
├── outputs/ # Processed data generated by the notebooks
│ ├── df_morts_par_drogue.csv
│ ├── df_drogue_par_age.csv
│ └── correlation_cocaine_crime.csv
│
├── PowerBI/
│ └── EscobAddictions.pbix # Interactive dashboard (3 pages)
│
└── img/
└── Cocaine_criminalité.png # Cocaine / crime correlation visualization
The file PowerBI/EscobAddictions.pbix contains three interactive pages:
- Deaths by drug type — breakdown by substance and sex (Connecticut, 2012–2018)
- Drug use by age group — percentage of users per drug and age bracket
- World map — cocaine-related death rates by country and year (1990–2019), with an interactive time filter
| Tool | Use |
|---|---|
| Python 3 | Data processing and analysis |
| pandas | Data manipulation and cleaning |
| numpy | Numerical computations |
| matplotlib | Static visualizations |
| Power BI | Interactive dashboard |
| Jupyter Notebook | Analysis environment |
| GitHub | Collaboration and version control |
| Dataset | Source |
|---|---|
| Accidental drug overdose deaths — Connecticut | Kaggle / Chief Medical Examiner |
| Drug use by age group — USA | Kaggle / FiveThirtyEight |
| Substance-specific mortality rates | Our World in Data |
| Intentional homicides by country | World Bank |
| Drug-related prison admissions | UNODC — CTS |
git clone https://github.com/Wabfall/escobaddictions.git
cd escobaddictions
pip install pandas numpy matplotlib jupyter
jupyter notebookOpen rapport.ipynb for a full project overview, or any of the numbered notebooks for a specific study.
Project completed as a team for course 8PRO408 at UQAC (Summer 2023).
| Name | Role |
|---|---|
| Etienne CHEVROLLIER | Analysis & visualization |
| Andonin COUSSEAU | Analysis & data sourcing |
| Khassan TASSOUEV | Analysis & data sourcing |
| Damien BALLET | Analysis & Power BI |