Nov2025_Ganji_Rajesh_Cities_OfTomorrow.ipynb
Github link:https://github.com/ganji-rajesh/portfolio/blob/main/projects_in_2025/fabric_data_days.ipynb
Here are 10 deep insights derived from the analysis:
1. Green Cover is the "Golden Variable" The single strongest predictor of a high urban_sustainability_score is green_cover_percentage (Correlation: +0.69). This impact dwarfs all other infrastructural factors. Increasing green spaces provides roughly 3x the sustainability benefit of improving public transport access in this specific dataset.
2. The Density-Sustainability Paradox Contrary to the popular "compact city" theory, population_density and building_density show zero correlation (~0.00) with the overall sustainability score. A dense city is not inherently more sustainable in this model; high-density areas are just as likely to score poorly as low-density ones if they lack green cover or renewable energy.
3. Zoning is Neutral There is no "best" land use type. The average sustainability score is virtually identical (~0.48) across Commercial, Residential, Industrial, and Green Space zones. This implies that sustainability is determined by how a zone is managed (e.g., solar panels on a factory, trees in a commercial district) rather than what the zone is designated for.
4. The "Invisible" Carbon Footprint carbon_footprint is a major negative driver of sustainability (Correlation: -0.34), but interestingly, it does not correlate with road_connectivity or building_density. This suggests that in this city, carbon emissions are likely driven by individual consumption or industrial processes that are independent of the physical urban layout.
5. Crime is a Standalone Issue crime_rate negatively impacts the sustainability score (Correlation: -0.21), but it has no relationship with population_density or avg_income. High-density neighborhoods are not statistically more dangerous, nor are wealthier neighborhoods statistically safer in this dataset.
6. Renewables Outperform Transport renewable_energy_usage is the second most powerful lever for sustainability (Correlation: +0.48). Transitioning to renewable energy sources has a significantly higher impact on the city's score than expanding public_transport_access (Correlation: +0.20).
7. Environmental Equity Exists There is no correlation between avg_income and air_quality_index or green_cover_percentage. This indicates a highly equitable urban model where environmental benefits (and burdens) are distributed evenly across rich and poor neighborhoods, avoiding the common "environmental justice" pitfalls seen in many real-world cities.
8. Disaster Risk is Location-Agnostic disaster_risk_index is a major drag on sustainability (Correlation: -0.35), but it does not correlate with land_use_type. Industrial zones are not inherently "riskier" than Residential zones in this dataset, suggesting disaster risk is likely modeled as a city-wide or random geographic factor rather than a zoning-specific one.
9. Air Quality is Disconnected from Local Traffic Surprisingly, air_quality_index does not correlate with road_connectivity or public_transport_access. This suggests that air quality in this dataset might be driven by regional factors (e.g., wind patterns, external pollution) rather than local traffic congestion.
10. The "Actionable" Formula Because the input variables (like density, income, transport) have very low correlation with each other, they can be manipulated independently. A planner can increase Green Cover and Renewable Usage to drastically boost the city's score without needing to engage in the difficult, long-term politics of changing zoning laws or population density.