Insights
Different parks do different jobs. Comparing all of them on a single Vitality score is a useful summary; it also hides almost everything interesting. These pages cut the data several ways and ask: what kind of park is this, and what would a useful intervention look like? Every chart and number here is computed from the live ETL pipeline.
Jacobs vs Wilderness
Urban integration = average of edge activation, connectivity, and enclosure. Natural comfort = canopy + ravine + water + tree density. Most parks lean strongly one way; balanced hybrids are rare.
Vitality Score distribution
How the overall Jacobs Vitality Score is spread across all 3,273 parks.
Typology distribution
Number of parks classified into each typology.
Natural Comfort by typology
Mean Natural Comfort within each typology. Ravines and Wilderness lead by design.
Enclosure by typology
Mean Enclosure within each typology. Civic Squares and Parkettes lead.
Pick a question
Jacobs vs Wilderness
Two-axis chart of every Toronto park: urban integration vs natural comfort.
Detected urban patterns
Auto-detected groupings: tower-shadow parks, ravine-edge disconnects, suburban successes.
By neighbourhood
Per-neighbourhood roll-ups: average scores, dominant typologies, paradox parks.
Contested parks
Where public sentiment disagrees with the model. Public-feedback driven.
Best-connected parks
Highest Connectivity scores: many streets, entrances, and transit nearby.
Most naturally comfortable
Highest Natural Comfort. Canopy, ravine, water.
Best-enclosed parks
Strongest mid-rise frontage, Jacobs-style 'eyes on the park'.
Most balanced parks
Parks that work across all five dimensions, not just one.
Connectivity vs Comfort gaps
The biggest splits between urban placement and ecological cooling.
Most Jacobsian parks
Highest urban integration with at least moderate comfort.
Most isolated parks
Low connectivity and few active edges.
Waterfront paradox
Beautiful waterfronts cut off from the city.
Ravine paradox
Ecological strength plus urban disconnection.
Most-reviewed parks
Top 50 parks by Google review count. Public attention, not score.
Visitor paradoxes
Where civic attention diverges from the model: high attention with low score, and the inverse.
Hidden gems
Tiny parks (<1.5 ha) with dense Google reviews and high-confidence place matches.
Place-match queue
Google candidates flagged for manual review or rejected automatically.
Event coverage
Curated programming feeds, sources, unmatched events, low-confidence matches.
Activity coverage
Per-typology and per-neighbourhood real-data signal coverage.
Disagreement
Where humans disagree with the model: contested parks, loved-despite-weak-metrics, overrated.
Human vs model
Scatter plots of perception (1 to 5) vs structural score (0 to 100) across four dimensions.
Beloved calibration
Does the model approximately align with collective intuition? Curated 'iconic' park list compared to cohort.
Strong overperformers
Parks beating the median score of their cohort by 5 points or more. Same typology, similar size, same ravine or waterfront status.
Strong underperformers
Parks scoring 5 points or more below the median of comparable parks. Often placeholder parcels or parks with decayed edges.
Quick previews
Sorted ascending; lowest by design.