Data Science (Hons.) student with a Minor in Finance at York University, based in Toronto. I like using data to understand real social and economic problems — especially questions about cities, housing, public health, and the systems people depend on.
I care about doing analysis honestly: choosing methods that fit the question, being clear about what the data can and can't say, and reporting results even when they're inconvenient.
Toronto Overdose Spatial Analysis — a geospatial study of suspected non-fatal opioid overdoses in Toronto shelters (2018–2025).
I mapped incidents to Toronto's 158 neighborhoods, then used spatial regression to test what actually predicts overdose incidence. The headline result is a careful one: most of the apparent geographic clustering is explained by low-income concentration, and what's left is driven by a handful of individual shelter addresses rather than any neighborhood-wide pattern — a finding I show and stress-test, including where the model fails.
Tools: Python, pandas, GeoPandas, PySAL (spatial regression), folium, geopy.
- Languages: Python, SQL
- Data & analysis: pandas, NumPy, GeoPandas, spatial statistics (Moran's I, spatial lag/error models)
- Visualization: matplotlib, folium
- Learning next: scikit-learn, predictive modeling, and deploying models as small web apps
- Adding a predictive machine-learning layer to the overdose project (forecasting neighborhood incidence)
- A new project on housing and homelessness in Toronto, building on the same geospatial approach
- GitHub: @ayokumo
- LinkedIn: Ayokunmi Lawal
- X: @ayokunm1
- Instagram: @ayokunm1
- Location: Toronto, Ontario