Compare commits

..
3 Commits
5 changed files with 167 additions and 152 deletions

No files matched your search

Binary file not shown.
+15 -12
View File
@@ -1,23 +1,26 @@
# PulseCast paper
# PulseCast / Cool City paper
## Published
## Branch `feat/cool-city` — event paper
This branch is the **Cool City** event report. `main.md` is the 4-page event paper and `CoolCity-Hackathon-Report.pdf` is the generated PDF. The prior staffing paper remains published at the DOI below; `PulseCast-*.pdf` and `main-extended.md` are the prior paper, not the Cool City event report. Do not merge this branch until the event review.
## Published (prior staffing paper)
**Zenodo:** https://zenodo.org/records/24217623
**DOI:** https://doi.org/10.5281/zenodo.24217623
**Concept DOI:** https://doi.org/10.5281/zenodo.24217622
Two versions of the paper, plus the research trail.
| File | What it is | Status |
|---|---|---|
| `PulseCast-Hackathon-Report.pdf` | Upload-ready 4-page A4 hackathon report | Final draft |
| `PulseCast-Extended-Paper.pdf` | Upload-ready 11-page A4 Zenodo paper | Final draft |
| `main.md` | Markdown source for the hackathon report | Final draft |
| `main-extended.md` | Markdown source for the extended paper | Final draft |
| `ZENODO.md` | Copy-paste Zenodo metadata and upload checklist | Ready |
| `references.md` | Canonical reference list and verification checklist | Draft |
| `OUTLINE.md` | Writing plan, claims/non-claims, figures/tables | Working document |
| `figures/` | Generated figures used by both versions | Final |
| `CoolCity-Hackathon-Report.pdf` | 4-page A4 Cool City event report | Final draft (branch) |
| `main.md` | Cool City event report source | Final draft (branch) |
| `PulseCast-Hackathon-Report.pdf` | Prior staffing event report | Published; do not edit |
| `PulseCast-Extended-Paper.pdf` | Prior staffing extended paper | Published; do not edit |
| `main-extended.md` | Prior staffing extended source | Cool City rewrite post-event/TBD |
| `ZENODO.md` | Prior paper metadata; Cool City metadata TBD | Working document |
| `references.md` | Prior paper reference list | Working document |
| `OUTLINE.md` | Prior writing plan | Working document |
| `figures/` | Figures for both papers | Mixed; `fig_cool_routes.png` is the Cool City route map |
| `research/` | Research notes (code audit, privacy law, sensing, heatmap, PETs, applications, data quality) | Private; do not upload |
| `build/` | HTML/CSS/PDF build artifacts | Working files |
Binary file not shown.

After

Width:  |  Height:  |  Size: 350 KiB

+108 -140
View File
@@ -1,214 +1,182 @@
# PulseCast: Turning a Noisy Footfall Index into a Weekly Staffing Plan
# Cool City: A Public Heat-Safe Route and Cooling-Gap Screening Layer for Heilbronn
**Author:** Noel Sabia (ORCID 0009-0008-3064-7573)
**Affiliation:** —
**Date:** 14 November 2026
**Artifact:** https://pulse.noelsabia.dev · **API:** https://api-pulse.noelsabia.dev · **Code:** to be released after the hackathon
**Prototype:** https://pulse.noelsabia.dev · **Cool City:** live at https://pulse.noelsabia.dev/cool (`/cool/stage`, `/cool/city`; image v1.3-cool, commit `d908c32`) · **DOI:** not assigned
## Abstract
Inner-city store managers must decide next week's shifts without knowing next week's footfall. We present PulseCast, a decision-first staffing copilot built on eight passive Bluetooth footfall counters in a ~580 m × 280 m cluster in Heilbronn, Germany. The system fuses five months of 15-minute device counts with Open-Meteo weather and Baden-Württemberg public holidays, trains a global gradient-boosted model on a time-based split, and converts the forecast into a weekly shift plan through a transparent staffing heuristic. On a 28-day holdout, the weather-aware 7-day model reaches MAE 3.32, WAPE 27.5%, and R² 0.777, versus 5.40, 44.7%, and 0.462 for a seasonal-naive baseline; the day-ahead model reaches 3.36, 27.8%, and 0.763. The plan is generated in 19 ms cold on v1.1 (contract < 2 s), and the deployed system handles 120/120 concurrent SSE clients with zero errors and a p95 event latency of 331 ms. A code-level audit of the public 2026 firmware commit e8be8bf shows a ~10-second scan, a 16-bit dedup key, and a 0.7 multiplier; the deployed 2025 binary may differ, and the data confirm the 0.7 scaling and per-wake export, not the scan duration or truncation. We therefore frame the result as a relative activity index that supports staffing decisions, not a city-wide people counter, and state what calibration and privacy work absolute claims would require.
Heat is a measurable public-health threat, and Heilbronn already knows where its cool places are, but not how a person reaches them. Cool City is a public heat-safe route built on the existing eight-sensor pulse. A weather-aware global model forecasts the relative footfall index, reaching MAE 3.32, WAPE 27.5%, and R² 0.777 on a 28-day holdout versus 5.40, 44.7%, and 0.462 for a seasonal-naive baseline. On a clear 35.0 °C day (14 August 2026), the route layer combines OpenStreetMap walkable ways, the city tree cadastre, and solar geometry to model three routes from Stadtgalerie to Marktplatz: shortest 359.5 m / 4.8 min, shaded 732.3 m / 9.8 min, and accessible 438.2 m / 5.8 min. The route UI is live at https://pulse.noelsabia.dev/cool (`/cool`, `/cool/stage`, `/cool/city`; image v1.3-cool). At 16:00 the accessible route is recommended for Ruth: 43.8% modeled shade, 3.3 direct-sun minutes, no ramp, and water and benches. The shaded route has a higher shade share (50.0%) but, being twice as long, more direct sun (4.9 min); shade share alone is misleading. A relative socio-spatial screening index ranks the eight 300 m catchments with weights 0.40 footfall, 0.30 shade deficit, 0.20 access deficit, and 0.10 vulnerability; Hagenbuchersee 2 ranks first (0.6471). Vulnerability is the 17.26% 65+ share of Heilbronn Innenstadt (31.12.2016) and is constant across all eight, so it does not differentiate them. Shade is modeled, the screening index is relative rather than a measured heat map, and the pulse is estimated devices, not people.
## 1. Introduction
## 1. Problem
A store manager on Kaiserstraße knows the weather forecast and the staff roster. She does not know how many people will walk past next Saturday. Raw counts are not a decision; the manager needs a forecast, an explicit staffing rule, and a way to test a scenario. A dashboard that shows history does not close that gap.
Heat is not a comfort issue. The Robert Koch Institute estimated about **16,000 heat-related deaths in Germany in 2026 through calendar week 37** (95% prediction interval 14,400–17,600; 19.3 per 100,000) [1]. That is a model estimate, not a count, and it is not a Heilbronn-specific figure. Heilbronn’s commissioned FITNAH-3D report gives a citywide physiological equivalent temperature (PET) range of **16.1–44.5 °C**, with unshaded sealed street space exceeding **41 °C PET** [2]. The report does not say that every inner-city street reaches 41 °C; the claim is about unshaded sealed surfaces.
The Schwarz Digits *Pulse of the City* challenge asks teams to turn privacy-preserving pedestrian counts from Heilbronn into predictive tooling for advertising, store locations, and staffing [1]. We chose staffing: eight fixed locations, five months of multi-sensor history, and a clear daily/weekly rhythm. Store-location ranking would need city-wide coverage that eight sensors in a 580 m cluster cannot provide; a city-wide heatmap would be a visualization, not a measurement.
The city already publishes **almost 60 cool places** on a map, with an address list [3]. That is a useful inventory, but it is not a routing or time-aware service: it tells a person where cool places are, not how to reach one at 16:00 with shade, water, benches, and step-free access. A local councillor argues that the map and care-home information do not reach alone-living older people, who carry the greatest risk; we quote that as her assertion, not as an evaluated finding [4]. The 2022–23 summer zones were discontinued, according to local reporting [4].
PulseCast makes three contributions. It treats the decision as the product: the forecast feeds a weekly shift plan. It uses a global, weather-aware gradient-boosted model with a time-based holdout and a seasonal-naive baseline, rather than a model the data cannot support. It is deployed as a three-service system on SKE with a cached fallback and a live scenario layer. It also reports a code-level audit showing that the published value is a short-window count of randomized BLE addresses scaled by 0.7, not a 15-minute count of people.
This is not a city that has done nothing. Heilbronn adopted a climate-adaptation concept in April 2023 with 31 measures across five fields [8] and published a 2025 spatial climate-adaptation concept [9]. The gap is narrower and more actionable: **the person-level route**. The city knows where the cool places are; it does not yet give a person a safe way there, hour by hour, with the access needs that determine whether the route is usable.
The report also serves the Schwarz Digits criteria: **problem and benefit** (a manager gets a plan, not a dashboard), **originality** (decision-first tooling plus a measurement audit), **feasibility and future** (live on SKE, 120-client load test, calibration roadmap), **reliability** (120/120 connections, zero errors, cached fallback), **understanding** (Sections 2.2 and 6 state what the index does and does not measure), **teamwork** (data/ML, Go API, Nuxt UI, and infra workstreams against a written contract, with human integration), **documentation** (code, contract, metrics, and this report), and **technology** (Vue/Nuxt + Go + Onyx + FastAPI on namespace-only SKE). For the final pitch, the technical evidence supports the 50% idea/implementation weight, the honest limits support the storytelling, and the calibration/privacy roadmap is the 20% vision.
The existing eight-sensor pulse is the evidence spine, but it is a **relative device index**, not a count of people. Cool City does not hide that. It uses the index to forecast relative demand and to screen relative cooling gaps, and it states the measurement limits wherever a decision depends on them. The contribution is the public decision layer over a noisy city index: three routes for a persona, a relative socio-spatial screening ranking for the city, and the retail staffing plan as one commercial tenant.
We do not claim a production people counter, a city-wide heatmap, annual seasonality, live sensor data, or a proven return on investment. We claim a working decision loop, a credible relative forecast, and an honest account of the measurement limits.
## 2. Data and measurement audit
## 2. Data
The delivered PAX CSV contains 192,464 rows: 96,232 density readings and 96,232 battery readings, from 12 device IDs. After merging the two `density-01` IDs and excluding three test sensors, eight usable locations remain in a ~580 m × 280 m inner-city cluster. The **12 June model window** (12 June–7 November 2026) contains **83,156 density values**; the **18 June coverage panel** contains **82,170 readings** and is used for coverage statistics.
### 2.1 The delivered PAX data
The public 2026 Code for Heilbronn `lorax` firmware (commit `e8be8bf`) implements a ~10-second passive BLE scan, a 16-bit dedup key, a 0.7 multiplier, and a 900-second sleep [11]. The deployed 2025 binary may differ. The data confirm the **0.7 scaling and per-wake export**, not the scan duration or the 16-bit truncation: every one of the 83,156 values in the 12 June model window is an exact multiple of 0.7; the reconstructed count has median 12, p95 71, max 418, and **2.70% of those values are zero**.
The dataset contains 192,464 CSV rows: 96,232 density readings and 96,232 battery readings. There are 12 device IDs; after merging the two `density-01` IDs and excluding three test sensors, eight usable locations remain in a ~580 m × 280 m cluster: `density-01`, `hn-goetzentuem-01`, `hn-hagenbuchersee-01`, `hn-hagenbuchersee-02`, `hn-marahaus-01`, `hn-neckarufer-01`, `hn-soleo-01`, and `hn-stadtgalerie-01`.
Coverage is uneven. In the 18 June panel, only **28.2%** of 15-minute slots have all eight sensors; the modal count is seven (4,117 slots). Aggregating to 30, 60, or 120 minutes raises complete coverage to 63.4%, 69.7%, and 71.0%. Across all 96,232 density rows, 5.18% are zero, 10.36% are <1, and 39.57% are <5; in the 12 June window the shares are 2.70%, 8.32%, and 35.86%. Gaps reach four to six days per sensor and are not random. The 12 June panel has median 8.4, p95 49.7, and p99 79.8; across all 96,232 rows the values are 7.7, 46.9, and 78.4.
The 12 June model window (12 June–7 November 2026) has a median cadence of 15.3 minutes; values range 0–292.6, with median 8.4, p95 49.7, p99 79.8. Across all 96,232 density rows, the median is 7.7, p95 46.9, p99 78.4 and the multi-sensor mean is ~14.4. In the 12 June model window, the inner city peaks 18:00–21:00 local, top 19:00; Saturday is busiest on average (mean 19.1) [2]. The sensors are LoRaWAN-connected PAX counters that passively sniff Bluetooth Low Energy (BLE) advertisements and transmit aggregated device counts [1, 2]. The counts are estimates, not a census: device penetration, Bluetooth settings, and the scanning design all matter.
The forecast is a global `HistGradientBoostingRegressor` trained on hourly means of the published index, with calendar, lag, Open-Meteo weather [12], and sensor-identity features. On the 28-day holdout, the production 7-day model reaches **MAE 3.32, WAPE 27.5%, and R² 0.777**; the day-ahead weather model reaches **3.36, 27.8%, and 0.763**; the seasonal-naive baseline is **5.40, 44.7%, and 0.462**. Weather is the main driver: removing it raises day-ahead MAE from 3.36 to 3.96. The holiday flag changes 3.96 to 3.84, a small effect. These are single-site, 28-day results.
### 2.2 What one value actually is
| Audit layer | Finding | Population / caveat |
|---|---|---|
| Sensor network | 8 usable locations, ~15 min cadence, ~580 m × 280 m | Inner-city cluster, not city-wide |
| Values | `0.7 × U`; 100% multiples of 0.7 | 12 June model window (83,156 values) |
| Coverage | 28.2% all-eight slots; 4,117 modal-seven slots | 18 June coverage panel (82,170 readings) |
| Zeros / low counts | 5.18% zero, 10.36% <1, 39.57% <5 | All 96,232 density rows |
| Model | MAE 3.32, WAPE 27.5%, R² 0.777 | 28-day holdout; production 7-day model |
| Unit | Estimated devices, not people; 0.7 factor unvalidated | The core caveat for every downstream layer |
A code-level audit of the public 2026 Code for Heilbronn `lorax` firmware (commit e8be8bf) changes the interpretation of the numbers [3, 4]. The node scans BLE passively for about 10 seconds, counts distinct locally administered addresses, truncates the last two address bytes into a 16-bit ID, multiplies by 0.7, and sleeps for 900 seconds. The deployed 2025 binary may differ; the data confirm the 0.7 scaling and per-wake export, not the scan duration or truncation.
*Table 1. The measurement audit. Cool City inherits every limit in this table; it does not turn the index into people or into temperature.*
In the public 2026 firmware, the published value is `0.7 × U`, where `U` is the number of distinct randomized BLE addresses seen in a ~10-second scan. The 15-minute cadence is the sleep cycle, not the integration window; the effective duty cycle is ~1.1% in that code. Every one of the 83,156 values in the **12 June model window** is a multiple of 0.7; the reconstructed count has median 12, p95 71, max 418, and 2.70% of those values are zero. A hardware replacement creates a scale break in `density-01`; the legacy and current series are not pooled [5].
## 3. Cool City
This is a relative activity index, not a count of people. A 30-second passer-by has a ~4.4% chance of overlapping a scan; a one-hour lingerer can be counted up to four times. The index conflates flow and dwell, cannot distinguish phones from watches/beacons/cars, and the people-per-device ratio is unknown. The 0.7 factor is a provisioning default, not a validated penetration correction.
### 3.1 A person, three routes, and a clock
> **Practical consequence.** The model predicts the index well enough to compare relative peaks and drive a staffing rule. It does not predict people, and it does not support absolute footfall or revenue claims without calibration.
The demo persona is **Ruth, 78, who lives alone**. She is a clearly labelled composite, not a real person; the prototype can later use a real local story with consent. Her trip is **Stadtgalerie → Marktplatz**: short, ordinary, and exactly the kind of trip that heat turns into a risk. Both endpoints and all street names are OSM-verified. The routes start at the OSM-tagged main entrance of Stadtgalerie on the east side (the official address point is on the west side) and end at the centre of Marktplatz.
### 2.3 Coverage, gaps, and context
Cool City offers three routes:
Coverage is uneven. In the **18 June coverage panel (82,170 readings)**, only 28.2% of 15-minute slots have all eight sensors; the modal count is seven (4,117 slots). Aggregating to 30, 60, or 120 minutes raises complete coverage to 63.4%, 69.7%, and 71.0% [5]. Across all **96,232 density rows**, 5.18% are zero, 10.36% are <1, and 39.57% are <5; in the **12 June model window** the shares are 2.70%, 8.32%, and 35.86%. The worst month falls to 23.3% coverage at one location in the 18 June panel. Gaps reach four to six days; the cause is undocumented, but missing periods differ systematically from full-coverage periods [2, 5]. Weather comes from Open-Meteo: historical archive for training, current/7-day forecast for the dashboard. Temperature and precipitation are the useful features; cloud and wind did not improve the short test. Baden-Württemberg holidays are a feature and badge; they added little and are context, not a large driver [2].
1. **Shortest** — the direct path via Deutschhofstraße, Fleiner Straße, Kirchbrunnenstraße, and Windgasse.
2. **Shaded** — the detour via the unnamed shaded tunnel at the Deutschhof, the Deutschhof courtyard, and Kiliansplatz.
3. **Accessible** — the step-free route via Sülmerstraße and Falkenstraße that avoids the Marktplatz tram ramp.
| Sensor | Readings | Mean | Median | p95 | Max | Max daily gap |
|---|---:|---:|---:|---:|---:|---:|
| hn-stadtgalerie-01 | 8,982 | 21.1 | 8.4 | 67.2 | 137.2 | 6 days |
| hn-neckarufer-01 | 9,527 | 19.2 | 11.9 | 61.6 | 255.5 | 5 days |
| hn-hagenbuchersee-02 | 10,099 | 19.3 | 11.2 | 64.4 | 200.9 | 4 days |
| hn-marahaus-01 | 9,254 | 13.7 | 9.1 | 38.5 | 194.6 | 6 days |
| hn-hagenbuchersee-01 | 12,345 | 13.0 | 6.3 | 49.7 | 292.6 | 4 days |
| hn-goetzentuem-01 | 9,952 | 11.1 | 8.4 | 29.4 | 210.7 | 5 days |
| hn-soleo-01 | 10,327 | 10.6 | 8.4 | 28.0 | 111.3 | 4 days |
| density-01 | 12,670 | 9.9 | 7.7 | 28.0 | 63.0 | 35 days |
The shaded route uses the real OSM tunnel between Eichgasse and the Deutschhof courtyard. OSM has no signed name for it, so the bundle calls it “the shaded passage” and marks the name unverified; “Deutschhof”, “Taubenbrunnen”, and “Kiliansplatz” are OSM-verified. The courtyard is a dead end in OSM—the east/south passage is `access=no` and the north-east gate is `locked=yes`—so the route enters and leaves through the same tunnel. That out-and-back is why it is about twice as long as the shortest option.
*Table 1. Per-sensor coverage and value statistics. Except for the `density-01` max-gap entry, statistics use the 12 June–7 November 2026 model window; the `density-01` 35-day gap spans the longer history and hardware replacement.*
A time slider at **10:00, 12:00, 14:00, and 16:00** changes the recommendation as the sun moves. Each route card shows modeled shade coverage, direct-sun minutes, distance, duration, step-free status, water stops, and bench stops. The shade values are **modeled**, not measured. The audience vote is a comprehension device, not a data source: the room picks a route for Ruth, and the stage shows the split next to the model’s recommendation. Most people take the shortest; the model recommends the accessible route on access grounds, not because it has the least sun.
![Figure 1. Sensor map.](figures/fig_sensor_map.png)
### 3.2 Route and shade model
*Figure 1. The eight sensor locations in the ~580 m × 280 m inner-city cluster. Measured points only; no city-wide surface.*
The route layer uses the OpenStreetMap walkable network (footways, pedestrian streets, paths, steps, crossings, buildings, trees, benches, drinking water, toilets, and wheelchair and kerb tags; ODbL) [10], the city tree cadastre from January 2026 (crown diameter; dl-de/by-2.0) [5], and solar position for the selected hour. `access=no`/`private` ways and locked gates are excluded; steps and `wheelchair=no` ways are forbidden for the accessible route. The bundle covers 854 main building footprints—6 have a `height` tag and 403 have `building:levels`, so **445 use a type default** (the 12 m default for `building=yes` alone covers 314)—and 801 trees: 795 cadastre trees plus 102 OSM trees, deduplicated within 6 m. Tree canopy height is modeled as `min(4.5 + 0.45 × crown_dm, 14) m` and radius as `crown_dm/2`; tree heights are not measured.
![Figure 2. Hourly and weekday profile.](figures/fig_hourly_profile.png)
Buildings are swept along the shadow vector `height / tan(altitude)`; tree canopies are modeled as horizontal discs offset by `canopy_height / tan(altitude)`. Routes are sampled every 5 m, and a sample counts as shaded if it falls inside a building or tree shadow or on a way tagged `covered=yes`/`tunnel=yes`. Walking speed is constant at **1.25 m/s**, and `direct_sun_min = duration_min × (1 − shade_pct/100)`. Shade is clear-sky geometry: cloud cover is reported but not applied. The model is not a temperature model—it reports shade coverage and direct-sun minutes, not “degrees cooler.”
*Figure 2. Mean index by hour and weekday across the eight sensors. The peak is 18:00–21:00 local; Saturday is busiest on average.*
| Route | Distance / duration | 16:00 modeled shade | Direct sun 16:00 | Detour vs shortest | Step-free | Water / benches |
|---|---|---:|---:|---:|---|---|
| Shortest | 359.5 m / 4.8 min | 43.8% | 2.7 min | 0 m | Yes, but 15% ramp | 2 / 8 |
| Shaded | 732.3 m / 9.8 min | 50.0% | 4.9 min | +372.8 m (+103.7%) | Yes, but 15% ramp | 2 / 9 |
| Accessible | 438.2 m / 5.8 min | 43.8% | 3.3 min | +78.7 m (+21.9%) | Yes, no ramp | 2 / 6 |
## 3. Method
*Table 2. Route cards on the clear hot day 2026-08-14 (max 35.0 °C; 0% cloud at all four hours). Shade is modeled, not measured. Water = public drinking-water points within 30 m; benches = OSM benches within 40 m.*
### 3.1 Preprocessing and features
![Figure 1. Three routes from Stadtgalerie to Marktplatz.](figures/fig_cool_routes.png)
We parse timestamps as UTC, convert to `Europe/Berlin`, handle daylight saving, treat the two `density-01` IDs as one logical location but do not pool the legacy integer series with the 0.7-scaled replacement (the model uses the current device only), exclude the three test sensors, and aggregate to hourly means. The target is the hourly mean of the published index, not people.
*Figure 1. The three routes on the OSM base network (14 November 2026 snapshot; ODbL). Grey = buildings and walkable ways, green dots = tree cadastre and OSM trees. The shaded route’s final approach overlaps the accessible route and is drawn dashed. The map is illustrative; the metrics come from the 5 m shade model.*
Features are calendar (hour, weekday, month, weekend, holiday), lags (24, 48, 168 hours, plus 24-hour and 7-day rolling means), temperature and precipitation from Open-Meteo, and sensor identity. All lags and rolling windows end at the forecast issue time; the time-based split prevents leakage.
At **10:00** the shortest route has 61.6% shade and 1.8 direct-sun minutes, the shaded route 61.5% and 3.8 minutes, and the accessible route 67.4% and 1.9 minutes. The accessible route has the highest shade share in the morning, but the shortest route still has slightly less total sun because it is 1.0 minute shorter. At **14:00** shade falls to 27.4% (shortest), 42.6% (shaded), and 30.3% (accessible), with 3.5, 5.6, and 4.0 direct-sun minutes. At **16:00** the shaded route has the highest shade share, 50.0%, but 4.9 direct-sun minutes.
### 3.2 Model and validation
**The shaded route is not the lowest total sun exposure.** It has the highest shade share at 14:00 and 16:00, but it is 2.04× the shortest distance, so its total direct-sun minutes are higher at all four times. The shortest route is lowest at all four times (ties with the accessible route at 12:00). The accessible route is recommended for Ruth because it avoids the **15% rolled-kerb ramp** at the Marktplatz tram crossing (`node/12946835801`, `kerb=rolled`, `ramp:wheelchair=yes`, `incline=15.0%`), which the shortest and shaded routes use. The accessible route enters Marktplatz from the north via Sülmerstraße/Falkenstraße and returns **no kerb or ramp nodes** in the accessibility audit. Against the shortest route it costs **+78.7 m (+21.9%), +1.0 min, and +0.6 direct-sun minutes at 16:00**, in exchange for no ramp, two water stops, and six benches. The last 34 m on Marktplatz is `surface=sett` (cobblestone): step-free, but not a smooth-surface guarantee. The recommendation is an access preference, not a claim that the route is coolest.
We train a global `HistGradientBoostingRegressor` across all eight sensors. Train 12 June–10 October 2026; test the last 28 days, 10 October–7 November. Baseline: seasonal naive, same hour one week earlier. We report MAE, WAPE, and R². The global model pools strength across sensors while keeping sensor identity; with eight series and five months, a deep sequence model is not justified.
### 3.3 Cooling-gap screening index
The model serves 15 min, 1 h, 48 h, and 7 d. The 15 min/1 h horizons are nowcast/replay because the CSV is static; 48 h/7 d are planning horizons, with 7 d the hero.
### 3.3 From forecast to shift plan
The ML model predicts the published index; a deterministic heuristic turns it into staff:
The city layer uses the eight sensor catchments, each a **300 m Euclidean radius** (282,743 m², about a 4-minute walk at 75 m/min) [6]. It computes a **relative socio-spatial screening index** as a weighted sum:
```text
forecast index per hour
→ calibration factor k (unvalidated)
→ visitors per staff per hour (service standard)
→ required staff = ceil(index × k / standard × safety factor)
→ minimum staff, shift length, opening hours
→ shift blocks + coverage score + estimated cost
Gap = 0.40 × footfall
+ 0.30 × shade_deficit
+ 0.20 × access_deficit
+ 0.10 × vulnerability
```
The assumptions are visible and adjustable: calibration factor `k`, visitors per staff hour, labor cost per hour, minimum staff, shift length, and safety factor. `k` converts index units to visitors and is **unvalidated**; setting `k = 1` would treat one index unit as one visitor, which the audit says is not justified. The manager can edit assumptions and regenerate the plan with a scenario. This is decision support, not an optimizer with proven ROI; we have no POS, sales, queue, or labor-cost data.
The footfall, shade, and access components are min–max normalized across the eight catchments; vulnerability is a raw 0–1 share. `footfall` is the published relative device index, scored as 0.4 × min–max(mean) + 0.6 × min–max(p95) because peak footfall coincides with peak heat; its unit, metric, and aggregation interval are **not documented by the API**, so the semantics are assumed. `shade_deficit` is 1 minus a shade score: 0.5 × min(1, building footprint share / 0.45) + 0.5 × min(1, tree canopy share / 0.25), with 0.45 and 0.25 as planning heuristics. This is a **footprint/canopy proxy**, not measured shade and not the route-level shadow model. `access_deficit` is 0.35 water + 0.20 bench + 0.25 toilet + 0.20 step-free, each normalized as `min(distance, 300 m) / 300 m`; step-free is the distance to the nearest `wheelchair=yes` POI or official accessible toilet, not a routed barrier-free path. `vulnerability` is the raw 65+ share for Heilbronn Innenstadt, **17.26% (5,667 / 32,834, 31.12.2016)**. It is **constant across all eight catchments**, so it adds 0.01726 to every score and does **not differentiate them**; the ranking is driven by footfall, shade, and access.
## 4. System
| Rank | Catchment | Gap score | Main driver |
|---:|---|---:|---|
| 1 | Hagenbuchersee 2 | 0.6471 | Footfall, shade, and access above median; p95 64.4, canopy 1.8%, water 236.5 m |
| 2 | Stadtgalerie | 0.6306 | Highest peak footfall (p95 67.2); building shade 42.4%, water 124.2 m |
| 3 | Neckarufer | 0.6069 | p95 61.6, canopy 2.5%, water 208.1 m |
| 4 | Hagenbuchersee 1 | 0.4979 | Canopy 1.2%, toilet 162.4 m, p95 49.7 |
| 5 | Marahaus | 0.3752 | Water 233.6 m, canopy 4.7%, p95 38.5 |
| 6 | Soleo | 0.3433 | Lowest footfall score (0.025); high shade and access deficits |
| 7 | Götzenturm | 0.2931 | Low footfall score (0.064) |
| 8 | Density 01 | 0.2660 | Lowest footfall (0.000); battery 11 |
PulseCast runs as three namespace-scoped Deployments on STACKIT SKE plus a Cloudflare Tunnel connector: **Nuxt 4 SSR + Onyx** for the server-rendered shell, map, charts, forecast, plan, and scenario controls; a **Go API** for sensor/series/forecast/staffing endpoints, SSE fan-out, scenario state, and a cached fallback; and a **FastAPI ML service** for `/predict` and `/simulate`, loading a model baked into the image at build time.
*Table 3. Top-gap ranking under the base weights 0.40 / 0.30 / 0.20 / 0.10. Ranks 1 and 2 differ by 0.0165 (2.5%) and are treated as co-priority. Vulnerability is constant across all eight, so the ordering is driven by footfall, shade, and access. This is a relative socio-spatial screening index, not a measured heat map.*
The namespace does not allow Ingress, Jobs, CronJobs, cert-manager, or external-dns. We expose the app through a Cloudflare Tunnel, terminate TLS at Cloudflare, and run the batch at build time. The model and fallback JSON are baked into the images; there is no database or PVC. If FastAPI is unavailable, Go serves the cached forecast and labels it. SSE carries scenario changes with a 15-second heartbeat; polling is the fallback. The plan is generated in 19 ms cold on v1.1; the contract is under two seconds.
The top-3 set—Hagenbuchersee 2 (0.6471), Stadtgalerie (0.6306), Neckarufer (0.6069)—is stable under equal, shade-heavy, and access-heavy weights; only ranks 2 and 3 swap under shade-heavy weighting. Shade deficits cluster in a narrow 0.500–0.651 band, so footfall and access drive most rank differences. The index is relative to these eight inner-city sensors only: all lie within ~585 m of each other, 300 m catchments overlap by up to **87.8%**, and there is no outer-district coverage. **All eight catchments are inner city; this is not a citywide neighbourhood ranking.** Citywide disability statistics (11,730 people in 2023; the city says “almost 12,000”) [7] are too coarse for catchment-level differentiation. The index is **not a measured temperature or heat map** and must not be presented as one.
The public app is at https://pulse.noelsabia.dev and the API at https://api-pulse.noelsabia.dev. The v1.1 images are 16.9 MB (api), 249 MB (web), and 952 MB (ml). In the v1.1 120-client Cloudflare k6 run, 120/120 SSE connections opened and closed with zero connection errors; p95 event latency 331 ms, p95 scenario latency 646 ms, presenter POST 49.60 ms, over 4m57.9s. The system runs non-root with resource limits and a read-only root filesystem. The public dashboard is read-only; `/control` uses basic auth. We store no PII or MAC addresses in the application. This is a hackathon deployment: one API replica, no SSO, no HA, no live ingestion. We document those gaps.
### 3.4 Retail staffing as one application
![Figure 3. System architecture.](figures/fig_architecture.png)
The existing staffing plan remains a separate module. The same forecast schedules store staff; a retailer pays for that module. In this paper it is a commercial tenant, not the hero. The live plan is currently a snapshot with unstable scenario/window behavior, so we do not report its plan numbers here; the module is described qualitatively until the plan window and scenarios are stabilized. The Cool City paper’s contribution is the public route and screening layer, not a retailer dashboard.
*Figure 3. Three services plus `cloudflared`: Nuxt SSR, Go API/SSE and cache, FastAPI ML. Build-time batch; cached fallback; no Ingress, cert-manager, or Jobs.*
## 4. Evaluation
![Figure 4. Plan screen.](figures/fig_plan.png)
**Route-level metrics.** The frozen bundle reports modeled shade coverage, direct-sun minutes, detour, duration, step-free status, water stops, and bench stops for each route at four times. At 16:00 the accessible route gives 43.8% shade and 3.3 direct-sun minutes for a 438.2 m / 5.8 min walk, +21.9% detour against the shortest route, with no ramp and two water and six bench stops. The shortest route is lower in total sun (2.7 minutes) but uses the 15% ramp. The shaded route is the longest (732.3 m, +103.7%) and has the most sun (4.9 minutes) despite having the highest shade share at 16:00 (50.0%). The route-level goals are therefore met in part: the accessible option meets the access and shade-share target, not the minimum-sun target; the shortest option meets minimum sun, not the no-ramp target. Reporting both numbers is the point. Step-free continuity is from OSM tags, not a field survey.
*Figure 4. The v1.1 plan screen for `scenario=sunny-saturday-market`, `standard=standard` (live API query, 14 Nov 2026; reconstructed for print): Saturday peak 5 staff, 172 scheduled staff-hours, €2,752 labour cost, 100% coverage, versus a €5,600 flat schedule. The live screen also includes a “why this plan” context card; the app is at https://pulse.noelsabia.dev.*
**City-level screening.** The eight catchments score from 0.6471 (Hagenbuchersee 2) to 0.2660 (Density 01). Ranks 1 and 2 are a co-priority (0.0165 apart, 2.5%), and the top-3 set is stable under the tested weight variants. The sensitivity analysis shows the screening is not decision-grade for individual positions 2–3 under shade-heavy weights, but the top-3 priority set is stable. This is a **relative socio-spatial screening index over eight inner-city catchments**, not a citywide neighbourhood ranking and not a measured heat map.
## 5. Evaluation
**System metrics.** The inherited SSE path was measured at **120/120 connections opened and closed, zero errors, p95 event 331 ms, p95 scenario 646 ms, presenter POST 49.60 ms, over 4m57.9s** on v1.1. The retail plan module’s cold generation is 19 ms; that number belongs to the retail module, not the route layer.
### 5.1 Forecast accuracy
**Live deployment.** The Cool City UI is live at **https://pulse.noelsabia.dev/cool** (commit `d908c32`, image v1.3-cool). Ten requests per endpoint from one presenter machine through Cloudflare gave `/cool` **median 291 ms, p95 431 ms**; `/cool/stage` **median 146 ms, p95 179 ms**; and `/cool/city` **median 140 ms, p95 165 ms**. The vote endpoint accepted **20/20 POSTs with 201**, with **median 100 ms and p95 382 ms**; the tally was exactly **6 shortest / 7 shaded / 7 accessible**; a duplicate session returned **200** and did not double-count; reset cleared to **0**. This is a **single-client, Cloudflare-mediated measurement, not a concurrency or load test**. The inherited SSE metrics above are a separate path and are not combined with these numbers.
Table 2 reports the 28-day holdout from the generated `metrics.json`. The weather-aware 7-day model reduces MAE by about 19% relative to the no-weather 7-day model (4.08 → 3.32) and by about 39% relative to seasonal naive (5.40 → 3.32). WAPE falls from 44.7% to 27.5%, and R² rises from 0.462 to 0.777. The day-ahead weather model reaches 3.36, 27.8%, and 0.763; the longer horizon is slightly stronger (3.32 vs 3.36 MAE) and remains well above the baseline.
**Comprehension test plan.** A small task-based test with older adults and caregivers will ask participants to (1) choose a route for Ruth at 16:00, (2) explain why it is cooler or more accessible, (3) compare the 10:00 and 16:00 recommendations, and (4) find water, benches, and step-free access. The sample, recruitment, and results are **TBD**; this is a plan, not a completed study.
| Model | MAE | WAPE | R² |
|---|---:|---:|---:|
| Seasonal naive (same hour last week) | 5.40 | 44.7% | 0.462 |
| Day-ahead, no weather | 3.96 | 32.8% | 0.706 |
| Day-ahead, + temperature/rain | 3.36 | 27.8% | 0.763 |
| 7-day-ahead, no weather | 4.08 | 33.8% | 0.697 |
| 7-day-ahead, + temperature/rain (production) | **3.32** | **27.5%** | **0.777** |
| Evaluation layer | Metrics | Status |
|---|---|---|
| Route | Shade 43.8–50.0%, direct sun 2.7–4.9 min at 16:00, detour 0 to +103.7%, stops, steps | Measured on frozen bundle |
| Screening | Score 0.2660–0.6471; top-3 stable; ranks 1–2 co-priority; vulnerability constant | Ranked; sensitivity run |
| System | SSE 120/120, p95 event 331 ms, p95 scenario 646 ms, presenter 49.60 ms | Measured on inherited v1.1 path |
| Live UI / vote | `/cool` 291/431 ms; `/cool/stage` 146/179 ms; `/cool/city` 140/165 ms (median/p95); vote 20/20 201, 100/382 ms, tally 6/7/7 | Measured; single client via Cloudflare |
| Comprehension | Task success, explanation quality, route preference | Plan only; TBD |
*Table 2. Model comparison on the last 28 days, from the generated `metrics.json` (`n = 5,350`). The mission brief’s 3.36/27.8/0.779 was the quick baseline; the canonical generated metrics are the production 7-day 3.32/27.5/0.777 and day-ahead 3.36/27.8/0.763. The target is the published index, not people.*
*Table 4. Evaluation status. Shade is modeled; the screening index is relative and socio-spatial; no temperature or causal claim is made.*
![Figure 5. Forecast versus actual.](figures/fig_forecast.png)
## 5. Privacy and limits
*Figure 5. Forecast versus actual for one sensor over the 28-day holdout. The model captures the daily/weekly shape but misses spikes; the band is a planning aid, not a calibrated 95% interval unless coverage is reported.*
Cool City uses aggregate device counts only. It does not store MAC addresses or PII, does not require accounts, and does not upload a visitor’s location. The audience vote is aggregate; a random session ID in local storage is enough. A production deployment needs a DPIA and a documented legal basis before collecting MAC-derived counts, with the city as controller and a legal review before launch. The existing index is estimated devices, not people; device penetration is unknown; the 0.7 factor is unvalidated; and only 28.2% of 15-minute slots in the 18 June panel have all eight sensors. The live latency figures come from a single presenter machine through Cloudflare, not a concurrency or load test.
### 5.2 Ablation and per-sensor behavior
The route layer has its own limits. Shade is modeled from buildings and tree crowns, not measured with radiometers. **445 of 854 main building footprints use type-default heights** (the 12 m default alone covers 314), tree heights are inferred from crown diameter, the model is 2D in plan view, and clear-sky geometry ignores clouds, awnings, facade reflections, and indoor shade. The bundle flags the **42.6% shade / 5.6 direct-sun minutes for the shaded route at 14:00** as its least-certain value. OSM is a 14 November 2026 snapshot, not an official city dataset; benches, water, toilets, and kerb tags may be incomplete; step-free status comes from tags, not a field survey; the accessible route’s last 34 m is cobblestone; and the Deutschhof courtyard is an out-and-back. The route metrics are suitable for comparing options, not for safety or medical advice.
Weather is the main driver. In the day-ahead ablation, removing it raises MAE 3.36 → 3.96. The holiday flag changes 3.96 → 3.84 (0.12); it is context, not a large lift. The production 7-day model is 3.32 with weather and 4.08 without. These are single-site, 28-day results.
The screening index has separate limits. Catchments are Euclidean 300 m circles, not network isochrones, and do not model the Neckar, rail, or major-road barriers. Footfall semantics—unit, metric, and aggregation interval—are assumed because the API does not document them. The shade component is a footprint/canopy-share proxy without building heights or solar geometry; it is not the route-level shadow model. Step-free access is proxied by distance to a `wheelchair=yes` POI, not a barrier-free route. Vulnerability is a **2016 district-level 65+ share, constant across all eight catchments**; it does not vary at catchment scale and does not differentiate them. All eight sensors sit within ~585 m and catchments overlap by up to 87.8%, so the screening is relative and local, not a citywide neighbourhood ranking. It does not establish that an intervention will cause a health improvement. We make no “4 °C cooler” claim and no causal claim for shade, water, or bench interventions. The current prototype is a decision-support layer over a noisy index, not a validated public-health instrument.
Per-sensor errors vary (Table 3): Stadtgalerie has the highest MAE (5.83); Hagenbuchersee-01 the highest WAPE (38.6%); density-01 the lowest MAE (2.26). The global model is pooled; local errors are not.
## 6. Future
| Sensor | MAE | WAPE | R² |
|---|---:|---:|---:|
| density-01 | 2.26 | 30.8% | 0.722 |
| hn-goetzentuem-01 | 2.40 | 26.9% | 0.718 |
| hn-hagenbuchersee-01 | 2.47 | 38.6% | 0.602 |
| hn-marahaus-01 | 2.67 | 29.8% | 0.671 |
| hn-soleo-01 | 3.19 | 20.2% | 0.548 |
| hn-neckarufer-01 | 3.63 | 28.2% | 0.695 |
| hn-hagenbuchersee-02 | 4.12 | 31.1% | 0.731 |
| hn-stadtgalerie-01 | 5.83 | 25.3% | 0.765 |
The next step is a city-owned public trust: the city controls the route and screening layer, the underlying data is open where licensing allows (tree cadastre, OSM, city cool places, aggregate population), and the retail module is a commercial tenant rather than the product’s owner. The route/shade bundle—3 routes, 12 route/time records, 35 segments, 92 place features, 87.5 KiB—and the screening weights are published with the paper, tested with older adults and caregivers, and transferred to a second city with a different street network. Field validation of modeled shade and a small route-choice study would turn the current design into evidence. The long-term vision is a public heat-safe wayfinding layer that any city can adopt, with the retail staffing module as one paying application.
*Table 3. Per-sensor holdout error for the production 7-day model (`n ≈ 669` per sensor).*
## 7. TBD register
### 5.3 System metrics and decision evidence
| Item | Why it is TBD | Next step |
|---|---|---|
| Comprehension test | Plan only; no participants yet | Recruit older adults/caregivers with the city; publish tasks and results |
| Field verification of tags and surfaces | Step-free status, kerb/ramp, water, bench, and toilet tags are OSM-based, not surveyed; the accessible route’s last 34 m is cobblestone | Walk all three routes; check the 15% ramp, the tunnel, and the toilet tags |
| Field validation of modeled shade | Clear-sky 2D geometry with default building heights and inferred tree heights | Radiometer transects on the same routes; compare modeled and measured direct sun |
| Production licence for city-map layers | The official cool-place, toilet, and water layers have no explicit open-data licence | Confirm permission with the city before production |
| Metric | Value |
|---|---:|
| Plan generation latency | 19 ms cold on v1.1 (contract < 2 s) |
| Concurrent SSE clients | 120/120 opened and closed; 0 connection errors |
| SSE p95 event latency | 331 ms |
| SSE p95 scenario latency | 646 ms |
| Presenter POST latency | 49.60 ms |
| Duration | 4m57.9s |
| Image sizes (api / web / ml) | 16.9 MB / 249 MB / 952 MB |
| Cold deployment time | not measured |
*Table 4. System metrics from the v1.1 deployment and the v1.1 120-client Cloudflare k6 run. Cold deployment time was not measured.*
The v1.1 120-client Cloudflare run is the measured result: 120/120 connections opened and closed, zero errors, p95 event 331 ms. The plan screen compares the recommended schedule to a flat schedule on coverage and cost under stated assumptions; this is a heuristic comparison, not an ROI claim.
### 5.4 What the evaluation does and does not show
The evaluation shows that a global, weather-aware model beats a seasonal-naive baseline on a time-based holdout in one cluster. It does not show that the model predicts people, generalizes to another city, that weather is causal, or that the plan improves profit. The 28-day test is short, weather is observed, and the target is a ~10-second BLE index. It is a relative decision-support signal, not a breakthrough.
## 6. Discussion and Limitations
The main limitation is the measurement. The sensor samples ~10 seconds out of every 15 minutes, counts only randomized BLE addresses, truncates to 16 bits, multiplies by 0.7, and cannot distinguish people from devices. The value is a relative index, not a census. Penetration is unknown; a single factor cannot absorb device ownership, multiple devices per person, Bluetooth settings, and the short scan window.
Coverage is also a limitation. Eight sensors in 580 m × 280 m cannot support a city-wide heatmap or store-location recommender. Only 28.2% of 15-minute slots have all eight sensors; gaps are not random. The 28-day holdout is short, weather is observed, and there is no event log, POS, or labor-cost data. The deployment has one API replica, no SSO/HA, and no live ingestion; 15 min/1 h are replay/nowcast.
**Privacy.** The app uses aggregate counts only, stores no MAC addresses or PII, and does not redistribute the raw challenge CSV. The sensing layer is upstream of this application. A production deployment would need a DPIA and a documented legal basis before collecting MAC-derived counts.
**What to keep:** decision-first framing, global model, time-based holdout, seasonal-naive baseline, transparent heuristic, cached fallback, build-time batch, honest limitations. **What to replace:** sensing (thermal/radar/camera/probe/manual), calibration (site-specific, with uncertainty), privacy (no raw MAC retention, central DP, k-threshold), spatial model (network-aware hierarchical field), forecast (quantile GBM, forecast weather, conformal calibration), platform (auth, observability, SLOs) — in that order. Fix the measurement before scaling the model.
## 7. Related Work
Passive Bluetooth/Wi-Fi sensing consistently finds a small, biased sample: counts track relative trends, absolute levels need local calibration [6, 7]. MAC randomization reduces linkability but does not make passive counting a census, and fingerprinting can defeat it [7]. Our audit is consistent and makes the sampling design explicit. Gradient-boosted trees remain a strong baseline for tabular, covariate-rich, small-data problems [8]. Weather affects pedestrian activity, but its magnitude is context-dependent [9]; we treat it as predictive, not causal.
## 8. Conclusion
PulseCast shows that a small, weather-aware global model can turn a noisy inner-city footfall index into a weekly staffing plan that beats a seasonal-naive baseline and runs in 19 ms cold on a real deployment. It also shows that the index is not a people counter: each value is a short BLE scan scaled by 0.7, sampled every 15 minutes, with unknown penetration and incomplete coverage. The honest conclusion is narrower and more useful than a city-scale claim: the decision loop works, the relative forecast is credible, and the next step is calibration and a privacy-first sensing layer. We would rather ship a plan a manager can use and state what it does not measure than promise a city-wide heatmap the data cannot support.
## Artifacts, data, and AI disclosure
The live app is https://pulse.noelsabia.dev (API: https://api-pulse.noelsabia.dev); the code repository will be released after the hackathon. The raw challenge CSV is not redistributed. Derived aggregates, the model card, and the metrics JSON can be shared subject to the organizers’ permission. The app stores no MACs or PII; the sensing caveats are in Section 2.2 and the privacy box is in Section 6.
AI tools were used for drafting and editing. The authors checked each number and citation; some sources are project documentation or estimates. AI tools are not authors and take no responsibility for the content.
*Table 5. Open items. These are deliberately not filled with plausible-looking numbers.*
## References
[1] Schwarz Digits / Code for Heilbronn. *The Pulse of the City*, challenge brief, 2026. Local: `OKF_Vault/Schwarz Digits/FCH Challenge The _Pulse of the City_.pdf`.
[1] Robert Koch-Institut. *Hitzemortalität in Deutschland 2026*. https://www.rki.de/DE/Themen/Gesundheit-und-Gesellschaft/Gesundheitliche-Einflussfaktoren-A-Z/H/Hitze/Bericht_Hitzemortalitaet.html
[2] PulseCast project. `OKF_Vault/mission/data.md`, `mission/mvp.md`, `mission/architecture.md`, 2026.
[2] Stadt Heilbronn. *FITNAH-3D-Abschlussbericht zur Raumstruktur- und Klimaanpassung*, 2025. https://www.agenda21-hn.de/wp-content/uploads/2025/07/034_2025_DS_034_2025_Anlage_5_Abschlussber_Raumstrukt_Klimaanpassung.pdf
[3] Code for Heilbronn. “Passantenzahlen messen mit LORA,” 18 September 2025. https://codeforheilbronn.de/blog/2025/09/passantenzahlen-messen-mit-lora/
[3] Stadt Heilbronn. *Gesundheit und Hitze: kühle Orte*. https://www.heilbronn.de/leben/gesundheit/staedtisches-gesundheitsamt/gesundheit-und-hitze.html
[4] Code for Heilbronn. `lorax` / `sensor-pax` firmware and backend, commit `e8be8bf5c11e1b76c88c3f176382c6df99eea989` (develop, 18 June 2026). https://codeberg.org/cfhn/lorax/commit/e8be8bf5c11e1b76c88c3f176382c6df99eea989; upstream `ESP32-Paxcounter` and `libpax`, https://github.com/cyberman54/ESP32-Paxcounter and https://github.com/dbinfrago/libpax.
[4] Rhein-Neckar-Zeitung. “Hitze in Heilbronn: und nicht genug Hilfe,” 2026. https://www.rnz.de/region/sinsheim-kraichgau_artikel,-Kampf-gegen-hohe-Temperaturen-Hitze-in-Heilbronn-und-nicht-genug-Hilfe-_arid,2401501.html
[5] PulseCast data-quality analysis, `paper/research/07-data-quality.md`, 2026.
[5] Stadt Heilbronn. *Baumkataster Januar 2026*, dl-de/by-2.0. https://opendata.heilbronn.de/dataset/baumkataster
[6] A. Kurkcu and K. Ozbay. “Estimating pedestrian densities, wait times, and flows with Wi-Fi and Bluetooth sensors.” *Transportation Research Record* 2644(1):72–82, 2017. DOI: 10.3141/2644-09.
[6] Stadt Heilbronn. *Bevölkerung nach Altersgruppen*, Planungsbezirke. https://opendata.heilbronn.de/dataset/bev%C3%B6lkerung-nach-altersgruppen-11
[7] J. Martin, T. Mayberry, C. Donahue, L. Foppe, L. Brown, C. Riggins, E. C. Rye, and D. Brown. “A study of MAC address randomization in mobile devices and when it fails.” *Proceedings on Privacy Enhancing Technologies* 2017(4):365–383. DOI: 10.1515/popets-2017-0054.
[7] Stadt Heilbronn. *Menschen mit Behinderung* (11,730 in 2023); city page “knapp 12.000.” https://opendata.heilbronn.de/dataset/menschen-mit-behinderung · https://www.heilbronn.de/leben/barrierefreiheit-inklusion-leben-mit-behinderung.html
[8] J. H. Friedman. “Greedy function approximation: a gradient boosting machine.” *Annals of Statistics* 29(5):1189–1232, 2001. DOI: 10.1214/aos/1013203451.
[8] Stadt Heilbronn. *Klimaschutzteilkonzept zur Anpassung an den Klimawandel*, adopted 27 April 2023, 31 measures. https://www.heilbronn.de/umwelt-mobilitaet/klima-klimaschutz/anpassung-an-den-klimawandel.html
[9] L. F. Miranda-Moreno and A. C. Lahti. “Temporal trends and the effect of weather on pedestrian volumes: a case study of Montreal, Canada.” *Transportation Research Part D* 22:54–59, 2013. DOI: 10.1016/j.trd.2013.02.008.
[9] Stadt Heilbronn. “Hitze im Blick: konkretisiertes Klimaanpassungskonzept,” 2025. https://www.heilbronn.de/kommunikation/pressemitteilungen/details/artikel/hitze-im-blick-konkretisiertes-klimaanpassungskonzept-zeigt-wege-fuer-besseres-stadtklima.html
[10] OpenStreetMap contributors. *OpenStreetMap*, ODbL. https://www.openstreetmap.org/copyright
[11] Code for Heilbronn. `lorax` / `sensor-pax` firmware, commit `e8be8bf5c11e1b76c88c3f176382c6df99eea989` (develop, 18 June 2026). https://codeberg.org/cfhn/lorax/commit/e8be8bf5c11e1b76c88c3f176382c6df99eea989
[12] Open-Meteo. *Historical and forecast weather APIs*, CC BY 4.0. https://open-meteo.com/
+44
View File
@@ -0,0 +1,44 @@
# Cool City — source-checked claims
Snapshot: 14 Nov 2026. This note separates verified claims from inferences. Use it before writing the paper, deck, or stage script.
## Verified claims
| Claim | Source | Exact support | Use |
|---|---|---|---|
| RKI estimated ~16,000 heat-related deaths in Germany in 2026 through calendar week 37 (95% PI 14,400–17,600; 19.3 per 100,000) | https://www.rki.de/DE/Themen/Gesundheit-und-Gesellschaft/Gesundheitliche-Einflussfaktoren-A-Z/H/Hitze/Bericht_Hitzemortalitaet.html | “Im Jahr 2026 wurden deutschlandweit bis zur Kalenderwoche 37 rund 16.000 [14.400; 17.600] hitzebedingte Sterbefälle geschätzt.” | Say “estimated”, “through KW37”, “Germany”, not a count and not Heilbronn. |
| Heilbronn’s commissioned FITNAH-3D report gives a citywide PET range of 16.1–44.5 °C; unshaded sealed street space can exceed 41 °C PET | https://www.agenda21-hn.de/wp-content/uploads/2025/07/034_2025_DS_034_2025_Anlage_5_Abschlussber_Raumstrukt_Klimaanpassung.pdf | “Die PET umfasst im Stadtgebiet Heilbronns eine Spannbreite von 28,4 °C und reicht von 16,1 - 44,5 °C.” / “Die höchsten Werte sind über versiegelten Gewerbegebieten, dem Straßenraum und Wohnsiedlungen ohne großkronigem Baumbestand zu finden (mehr als 41 °C PET; extreme Wärmebelastung).” | Say “commissioned model”, “unshaded sealed street space”, not “every inner-city street”. |
| Heilbronn lists almost 60 cool places on a map and an address list | https://www.heilbronn.de/leben/gesundheit/staedtisches-gesundheitsamt/gesundheit-und-hitze.html | “Hier werden fast 60 kühle Orte im Heilbronner Stadtgebiet gelistet…” / “Adressliste der Karte für heiße Tage als Text” | Say “map + address list, not a routing or time-aware service.” |
| The 2022–23 summer zones were discontinued | https://www.rnz.de/region/sinsheim-kraichgau_artikel,-Kampf-gegen-hohe-Temperaturen-Hitze-in-Heilbronn-und-nicht-genug-Hilfe-_arid,2401501.html | “Sommerzonen waren 2022 und 2023 in mehreren Straßen in Heilbronn eingerichtet. Inzwischen ist das Angebot eingestellt.” | Newspaper source; say “reported”. |
| A councillor says the cool map and care-home information do not reach alone-living older people | https://www.rnz.de/region/sinsheim-kraichgau_artikel,-Kampf-gegen-hohe-Temperaturen-Hitze-in-Heilbronn-und-nicht-genug-Hilfe-_arid,2401501.html | Stadträtin Isabell Steidel: “Wir haben eine Karte mit kühlen Orten … und Informationen an Pflegeeinrichtungen – beides erreicht die alleinlebenden Hochbetagten nicht, die das größte Risiko tragen.” | Quote as a councillor’s assertion, not an evaluated fact. |
| City tree cadastre: January 2026, UTM32N shapefile, dl-de/by-2.0, crown diameter | https://opendata.heilbronn.de/dataset/baumkataster | “Baumkataster Januar 2026 \| UTM32N shapefile”; “Lizenz dl-de/by-2-0”; “Stammumfang und Kronendurchmesser” | Good for shade modeling. Check the WGS84 entry date caveat. |
| Klimawäldchen: 72 trees, 2,245 shrubs, 13 mist columns; the city reports up to 11 K difference | https://www.swr.de/swraktuell/baden-wuerttemberg/heilbronn/hitze-abkuehlung-wassernebel-heilbronn-experimenta-gaensmarkt-klimawandel-100.html | “Insgesamt sind dort 72 Bäume und über 2.245 Stauden gepflanzt worden. Auch 13 Wasser-Nebelsäulen gibt es hier. … teils Temperaturunterschiede von bis zu 11 Grad erreicht werden, heißt es.” | Attribute the 11 K to the city (“heißt es”), not an independent measurement. |
| Population by age is available at planning-district level | https://opendata.heilbronn.de/dataset/bev%C3%B6lkerung-nach-altersgruppen-11 | “Daten werden auf Ebene der Planungsbezirke dargestellt.” | Aggregate only; not block-level. |
| Disability statistics: 11,730 (2023); city says “knapp 12.000” | https://opendata.heilbronn.de/dataset/menschen-mit-behinderung · https://www.heilbronn.de/leben/barrierefreiheit-inklusion-leben-mit-behinderung.html | CSV: “Insgesamt ;11.895;11.780;11.730” (2019/2021/2023). City: “In Heilbronn leben aktuell knapp 12.000 Menschen mit einer anerkannten Schwerbehinderung.” | Citywide, not small-area; not all heat-vulnerable. |
| Climate-adaptation concept: adopted 27 April 2023, 31 measures, five fields | https://www.heilbronn.de/umwelt-mobilitaet/klima-klimaschutz/anpassung-an-den-klimawandel.html | “Das Klimaschutzteilkonzept zur Anpassung an den Klimawandel wurde vom Gemeinderat am 27.04.2023 beschlossen und umfasst 31 Maßnahmen aus fünf Handlungsfeldern.” | Shows the city is acting; do not say “nothing exists”. |
| 2025 spatial climate-adaptation concept; hotspot Hauptbahnhof / Allee / Mannheimer Straße / Rollwagstraße; temporary shading at Marktplatz, Hafenmarkt, Berliner Platz | https://www.heilbronn.de/kommunikation/pressemitteilungen/details/artikel/hitze-im-blick-konkretisiertes-klimaanpassungskonzept-zeigt-wege-fuer-besseres-stadtklima.html | “Das Gebiet zwischen Hauptbahnhof, Allee, Mannheimer Straße und Rollwagstraße wurde in dem Konzept … vertieft ausgearbeitet.” / “Anbringung temporärer Verschattungselemente … Marktplatz, Hafenmarkt und Berliner Platz.” | Good for the city-gap module. |
| MapLibre is already a dependency; the current sensor map is schematic | Repo: `web-nuxt.git/main/package.json`; `app/components/SensorClusterMap.vue` | `"maplibre-gl": "6.18.0"`; “Schematic”; “Relative sensor positions … Not a city-wide heatmap.” | Reuse the dependency; do not overclaim map accuracy. |
## Claims to avoid or correct
- **“Static list”** — the city publishes a map and an address list. Say “map + address list, not a routing or time-aware service.”
- **“No standalone heat action plan”** — a 2025 newspaper headline says none; the 2026 status is not confirmed by a current primary city document. Do not claim. Say “the city has a 2021 climate-adaptation concept (adopted 2023, 31 measures), a 2025 spatial concept, and a 2023 heat-health action package; what it lacks is a person-level route.”
- **“Five-point plan”** — sources conflict (five vs six). Say “2023 heat-health action package” without a count.
- **“Inner city above 41 °C”** — the source supports unshaded sealed street space, not every inner-city street. Say “the commissioned model shows unshaded sealed street space above 41 °C PET.”
- **“4 °C cooler”** — never say. Use modeled shade coverage and direct-sun minutes. A temperature claim needs a validated radiant-temperature model and field data.
- **“Ruth, 78, lives alone”** — a composite persona, not a real person. Label it as illustrative, or use a real local story with consent.
- **Live staffing numbers** — base, sunny and rainy currently all return 162 staff-hours / €2,592; the plan window is 15–20 Nov while the page says 15–21. Treat as a live snapshot; do not put these numbers on the Cool City stage.
- **OSM counts** — our Overpass snapshot on 14 Nov 2026, not an official dataset. Label the query and the coverage gap.
## Data sources for the build
- City tree cadastre: https://opendata.heilbronn.de/dataset/baumkataster (dl-de/by-2.0)
- City cool places: https://www.heilbronn.de/leben/gesundheit/staedtisches-gesundheitsamt/gesundheit-und-hitze.html (almost 60 places; map + address list)
- City heat/climate reports: the FITNAH-3D report and the 2025 spatial concept above
- OSM: buildings, trees, benches, drinking water, toilets, wheelchair tags (ODbL)
- Open-Meteo: historical and forecast weather (CC BY 4.0)
- Population by planning district: https://opendata.heilbronn.de/dataset/bev%C3%B6lkerung-nach-altersgruppen-11
## Paper angle
Use the RKI estimate and the city’s own reports to establish the public problem, then make the narrow claim: the city knows **where** the cool places are, but not **how** to reach them with shade, water, benches and step-free access at a given hour. Cool City is the missing person-level layer. The retail staffing plan remains the commercial proof, not the hero.