Detected urban patterns

Auto-detected groupings that don’t fall out of typology or cluster, but matter to the experience of using a park. Each pattern is defined by a small rule set on the metrics and tells you something the typology alone won’t.

Excellent enclosure, poor comfort

Tightly framed by mid-rise but exposed to heat: strong urban form, weak ecological function. Common in dense pre-war neighbourhoods where the original tree canopy was lost.

Park	Typology	Edge	Conn	Comfort	Encl	Overall
Belmont Parkette	Urban Plaza	9	56	34	98	41
Bernard Avenue Road Allowance	Urban Plaza	81	47	32	98	57
Roxborough - Yonge St Traffic Island	Urban Plaza	50	66	24	98	43
Montclair Avenue Parkette	Urban Plaza	70	50	24	98	54
Parkview Gardens Parkette	Urban Plaza	0	44	26	98	28
Boswell Parkette	Urban Plaza	52	58	22	97	48
Glasgow Street Parkette	Urban Plaza	42	55	31	97	46
Jesse Ketchum Park	Urban Plaza	9	59	23	97	40
Honest Ed’s Park	Urban Plaza	31	48	31	97	44
Lillian H. Smith Park	Urban Plaza	63	40	32	97	47
Spadina Road Green	Urban Plaza	0	45	33	97	34
Spadina Road Green	Urban Plaza	17	25	24	97	33

Highly connected but socially weak

Many streets and transit stops nearby, but the surrounding land use is hostile or quiet. Connectivity does not, on its own, make a park work as a public place.

Park	Typology	Conn	Comfort	Encl	Overall
Sorauren Avenue Park	Neighbourhood Park	80	53	93	52
Dovercourt Park	Neighbourhood Park	70	57	79	52
Wychwood Barns Park	Neighbourhood Park	75	51	82	51
Jean Sibelius Square	Civic Square	74	51	96	51
High Park	Waterfront Park	89	75	89	51
Joseph Bannon Park	Athletic / Recreation Park	76	63	95	51
Mccormick Park	Neighbourhood Park	82	54	76	50
Dentonia Park	Neighbourhood Park	80	48	80	50
Wadsworth Park	Neighbourhood Park	78	56	76	50
Goulding Park	Neighbourhood Park	77	46	67	49
Lee Lifeson Art Park	Parkette	72	68	76	49
Monarch Park	Neighbourhood Park	79	62	80	49

Trapped by border vacuums

High Border-Vacuum Risk regardless of the park's other strengths: adjacent highways, rail, or institutional walls dominate the experience.

Park	Typology	Edge	Conn	Comfort	Encl	Overall
High Park	Waterfront Park	0	89	75	89	51
Christie Pits Park	Ravine / Naturalized Park	0	86	61	80	48
East Lynn Park	Neighbourhood Park	32	69	56	72	47
Earlscourt Park	Neighbourhood Park	0	82	58	74	46
Rennie Park	Waterfront Park	0	80	79	76	46
Cedarvale Park	Ravine / Naturalized Park	0	83	66	79	46
Ramsden Park	Ravine / Naturalized Park	0	84	52	94	46
Earl Bales Park	Waterfront Park	0	77	79	67	46
Snider Parkette	Urban Plaza	41	59	51	93	46
David A. Balfour Park	Ravine / Naturalized Park	0	85	71	86	45
Lawrence Park Ravine	Ravine / Naturalized Park	0	80	68	85	45
G. Ross Lord Park	Waterfront Park	0	78	75	63	45

Unexpectedly successful suburban parks

Suburban parks (low edge frontage typical of post-war street patterns) that still score well overall: design or programming pulling against the grain of the surrounding fabric.

Park	Typology	Edge	Conn	Comfort	Encl	Overall
Toronto Islands - Island Park	Waterfront Park	71	39	52	33	57
Toronto Islands - Centre Island Park	Waterfront Park	54	39	54	26	55

Overlooked high-performing parkettes

Tiny parks (under 0.5 ha) that punch well above their weight on enclosure and edge activation. Easy to dismiss on a map; central to neighbourhood life on the ground.

Park	Typology	Edge	Conn	Comfort	Encl	Overall
Market Lane Park	Urban Plaza	90	66	53	90	67
Maple Leaf Forever Park	Urban Plaza	81	71	37	83	65
Manor Community Green	Urban Plaza	88	46	51	84	64
Graham Park	Urban Plaza	80	65	38	81	63
Sonya'S Park	Urban Plaza	84	65	32	89	63
Norman Jewison Park	Urban Plaza	76	81	33	80	63
Grace - College Parkette	Corridor / Linear Park	80	68	38	91	62
James Canning Gardens	Urban Plaza	68	75	37	81	62
Ryerson Community Park	Urban Plaza	88	58	49	88	62
Asquith Green Park	Urban Plaza	68	77	33	80	61
Malta Park	Urban Plaza	61	68	53	84	61
Dundas - Watkinson Parkette	Urban Plaza	47	71	55	97	61

Tower-shadow parks

High enclosure score driven by tower walls rather than mid-rise frontage: passive surveillance is plentiful, but human-scale comfort suffers (wind, shadow, sterile ground floors).

Park	Typology	Edge	Conn	Comfort	Encl	Overall
Belmont Parkette	Urban Plaza	9	56	34	98	41
Roxborough - Yonge St Traffic Island	Urban Plaza	50	66	24	98	43
Montclair Avenue Parkette	Urban Plaza	70	50	24	98	54
Boswell Parkette	Urban Plaza	52	58	22	97	48
Glasgow Street Parkette	Urban Plaza	42	55	31	97	46
Jesse Ketchum Park	Urban Plaza	9	59	23	97	40
1365 Yonge	Urban Plaza	0	58	25	96	25
Bleecker Parkette	Urban Plaza	0	54	35	96	36
Gwendolyn Macewen Parkette	Urban Plaza	0	68	32	96	42
Mccaul - Orde Parkette	Urban Plaza	0	68	33	96	39
FOREST HILL PUBLIC LIBRARY - Building Grounds	Ravine / Naturalized Park	0	62	34	96	29
HARRISON POOL - Building Grounds	Urban Plaza	56	65	22	95	53

Ravine-edge disconnect

Ravine and naturalised parks scoring high on natural comfort while their edges are functionally sealed off from the city, so the ecological asset isn't reaching the people who live nearby.

Park	Typology	Edge	Conn	Comfort	Encl	Overall
City Wide Open Space	Ravine / Naturalized Park	0	23	92	57	34
Home Smith Park	Ravine / Naturalized Park	0	39	91	62	37
Trca Lands ( 60)	Ravine / Naturalized Park	0	27	91	59	35
City Wide Open Space	Ravine / Naturalized Park	14	40	91	89	44
City Wide Open Space	Ravine / Naturalized Park	14	40	91	89	44
Highland Creek Tributary	Ravine / Naturalized Park	0	26	91	68	36
Lower Don Parklands	Ravine / Naturalized Park	0	9	91	50	30
Rouge Park	Ravine / Naturalized Park	0	9	91	50	27
City Wide Open Space	Ravine / Naturalized Park	0	38	91	88	40
Moccasin Trail Park	Ravine / Naturalized Park	0	34	91	74	38
Trca Lands ( 70)	Ravine / Naturalized Park	0	29	91	66	35
City Wide Open Space	Ravine / Naturalized Park	0	32	91	78	34

Why these patterns?

Each pattern picks at a tension the headline score papers over. A park can be high-scoring and still be a tower-shadow plaza, a heat sink, or a ravine that nobody can reach.

The patterns are deliberately small and rule-based. If you disagree with a threshold, the relevant code lives in src/lib/insights.ts in the urbanPatterns()function. Change the numbers, rebuild the cache, and the page updates.