Internal · Data Quality
Signals & Activity Coverage
Every location activity on Glade is backed by one or more signals — species records, priority habitat classifications, or derived facility data. This page shows the distribution, quality, and coverage of those signals across the dataset.
Total signals
2,364,677
across all sources and activities
Locations covered
400,174
unique locations with ≥1 signal
Avg confidence
76.0%
mean across all signal rows
High confidence
65%
signals scoring ≥0.75
Signals by source
Each source contributes a different type of evidence. NBN species records tend to be the most granular; PHI habitat signals the most reliable at high confidence.
Natural England Priority Habitat Inventory — deciduous woodland, reedbeds, heathland and more. · 317,188 locations
Species records from the National Biodiversity Network — birds, mammals, insects, plants, fungi. · 180,749 locations
Signals computed from location type, facilities, and trail data (e.g. playgrounds, cycling). · 183,250 locations
Foraging-relevant plant and fungi species extracted from NBN Atlas using curated species lists. · 74,495 locations
Signal types & seasons
Signal type tells us what kind of evidence backs the activity. Season distribution shows how well the dataset captures time-of-year variation.
By type
Peak seasons
Year-round signals have no seasonal restriction. Season-tagged signals are shown only when the season matches (e.g. autumn fungi).
Confidence score distribution
Confidence runs from 0 (no evidence) to 1.0 (certainty). Most signals should cluster between 0.65–0.85 — enough evidence to show the activity, not so high as to be overclaiming. Currently 65% of signals score ≥0.75. A heavy left tail (below 0.60) would suggest weak base signals that may cause noise; a spike near the cap (0.90+) is normal for high-quality PHI habitat signals.
< 0.65
Weak
May produce low-relevance results
0.65 – 0.80
Good
Core working range for most activities
> 0.80
Strong
High-quality habitat or dense species records
Activity coverage
Coverage tells us which activities have the broadest signal support. Thin coverage (few locations) or low average confidence signals an activity that may show inconsistently in search results.
| Activity | Locations | Signals | Avg conf · distribution |
|---|---|---|---|
Wildlife | 398,554 | 605,123 | 80% |
Bird watching | 391,651 | 490,717 | 82% |
Foraging | 336,333 | 375,494 | 71% |
Insect spotting | 205,579 | 224,922 | 78% |
Seasonal nature | 145,463 | 145,463 | 71% |
Dog walks | 120,979 | 120,979 | 78% |
Heritage | 89,083 | 89,083 | 69% |
Forest walks | 81,944 | 81,944 | 63% |
Hiking | 44,050 | 44,146 | 64% |
Cycling | 39,974 | 40,020 | 67% |
Family friendly | 35,796 | 36,003 | 68% |
Picnic day out | 34,303 | 34,676 | 62% |
Playgrounds | 32,162 | 32,162 | 84% |
Mushroom foraging | 21,159 | 21,159 | 78% |
Waterside walks | 9,795 | 10,146 | 72% |
Mountain biking | 6,505 | 6,554 | 63% |
Farm visits | 3,497 | 3,585 | 80% |
Horse riding | 2,186 | 2,188 | 75% |
Easy access | 228 | 228 | 67% |
Trail running | 85 | 85 | 71% |
Static data snapshot - May 2026