feat: add OKF knowledge bundle for hackathon challenges

Capture all six Future City Hackathon 2026 challenges as OKF v0.2 concepts, with dataset and organizer concepts, indexes, and a log.

Configure release-please with the simple strategy so the first release PR bumps the repository from 0.0.0 to 0.1.0.
This commit is contained in:
Noel Sabia committed 2026-11-13 18:23:00 +01:00
commit 7b9893a89e
45 files changed
+45531

No files matched your search

@@ -0,0 +1,89 @@
---
type: Hackathon Challenge
title: "Measure where it matters: sensor placement for accurate heat maps"
description: Design an algorithm that places stationary heat sensors and plans mobile measurement routes so a coarse satellite map becomes a meaningful fine-grained heat map.
resource: ../../AI%20TRAQC/AI-TRAQC_Challenge_Measure-where-it-matters_Click_EN_v2.pptx
tags: [hackathon, future-city-hackathon-2026, ai-traqc, heat-mapping, sensor-placement, remote-sensing, optimization]
status: stable
generated: { by: opencode/deepseek-v4.1-flash, at: 2026-11-13T17:13:56Z }
sources:
- id: ai-traqc-deck
resource: ../../AI%20TRAQC/AI-TRAQC_Challenge_Measure-where-it-matters_Click_EN_v2.pptx
title: "Measure where it matters — AI TRAQC challenge deck (including speaker notes)"
author: team:ai-traqc
---
# At a glance
| Field | Value |
|---|---|
| Organizer | [AI TRAQC](/organizers/ai-traqc.md) |
| Core question | Where, when, and at how many locations must we measure for a city heat map to be meaningful? |
| Output | Sensor positions, mobile-unit routes with timing, and the number of measurement locations |
| Input | Open satellite data and open city data (NASA/USGS, ESA/Copernicus) |
| Bonus | Show that the approach also works for humidity or noise |
# The problem
Cities do not heat evenly. A river, a park, dense residential streets, and an industrial area can all behave differently on the same hot day. The available satellite data forces a trade-off between resolution and repetition:
- **Sentinel-3** provides a land surface temperature map every day, but at roughly 1 km per pixel. Streets and squares disappear inside a few pixels.[^ai-traqc-deck]
- **Landsat 8 and 9** provide finer images at roughly 100 m, but only every eight days, only around midday, and only when the sky is clear.[^ai-traqc-deck]
- Measuring everywhere with ground sensors and mobile units is too expensive because sensors and robot time are scarce.[^ai-traqc-deck]
The challenge is to decide where limited measurements buy the most information.
# Your task
Design an algorithm that starts **from scratch, without existing sensors**, and outputs:[^ai-traqc-deck]
1. **Where** stationary sensors are placed.
2. **Which routes** mobile units such as the robot dog take, and **when**.
3. **How many locations** need to be measured at all.
Example constraints are the number of sensors (for example `5+5` or `10+2`), the mission time, and priority areas such as day-care centres or care homes. The deck calls these examples, not fixed rules.
## Suggested workflow
1. **Choose placements from coarse data only.** Derive the placement from the Sentinel-3 baseline (1 km) and open city data. Do not peek at the fine reference while choosing locations.
2. **Optionally segment the city into climate zones.** Areas with similar buildings and surfaces behave similarly. The global Local Climate Zone map and OpenStreetMap are suggested open inputs.[^ai-traqc-deck]
3. **Simulate measurements.** Sample values at the chosen locations from the high-resolution Landsat reference (100 m, same day). The simulated readings correspond to a downward-facing infrared sensor and represent land surface temperature, not air temperature.[^ai-traqc-deck]
4. **Reconstruct a finer map.** Combine the simulated readings with the coarse baseline to produce a map on a finer grid, for example 100 m.
5. **Test against the reference.** Compare the reconstructed map with the full Landsat image. A curve of map error versus number of measurement points is one way to justify how many locations are needed.[^ai-traqc-deck]
## Timing
Landsat and Sentinel-3 are snapshots around midday, so the "when" needs a separate time axis. The speaker notes suggest ECOSTRESS scenes at different times of day or a modelled diurnal cycle. ECOSTRESS only covers up to about 52 degrees north; Sentinel-3 and Landsat cover every city.[^ai-traqc-deck]
# What you get
- Open satellite data from NASA/USGS and ESA/Copernicus, including Sentinel-3, Landsat 8/9, and optionally ECOSTRESS.
- An example city layout with a river and park, residential areas, a dense city centre, and an industrial area.
- Suggested open data: the global Local Climate Zone map (about 100 m) and OpenStreetMap for the path network.
- Suggested starting points from the deck: optimal sensor placement and informative path planning.
No challenge dataset is stored in this repository. The deck describes the data sources and the simulation workflow rather than shipping a file.
# Expected result
The deck is explicit that the result is **positions and routes, not just a map**: the algorithm itself should transfer to any city. The corrected heat map is the proof of how good the placement is, not the only deliverable.[^ai-traqc-deck]
# What makes a strong solution
- Placement uses only the coarse baseline and open data, then is evaluated honestly against the fine reference.
- The number of measurement locations is justified quantitatively, for example with an error-versus-count curve.
- The approach is transferable beyond the example city.
- Assumptions, limitations, and uncertainty are visible.
- Bonus: the same approach works for humidity or noise.
# Open questions
- How should priority areas be weighted against overall map error?
- How much does the time of day change the optimal placement?
- How should the algorithm handle clouds or missing satellite scenes?
# Sources
The challenge deck, including its detailed speaker notes, is the primary source for this concept.[^ai-traqc-deck]
[^ai-traqc-deck]: AI TRAQC, "Measure where it matters — sensor placement for accurate heat maps", challenge deck and speaker notes, 2026.
+142
View File
@@ -0,0 +1,142 @@
---
type: Hackathon Challenge
title: "Learning Guide: one place for Heilbronn's learning opportunities"
description: Build a prototype that consolidates and presents learning and leisure offers for children, young people, and parents in Heilbronn, with low maintenance and room to grow.
resource: ../../aim/aim_Challenge_Learning_Guide_readme.pdf
tags: [hackathon, future-city-hackathon-2026, aim, education, families, children, search, accessibility, privacy]
status: stable
generated: { by: opencode/deepseek-v4.1-flash, at: 2026-11-13T17:13:56Z }
sources:
- id: aim-readme
resource: ../../aim/aim_Challenge_Learning_Guide_readme.pdf
title: "aim Challenge Learning Guide — README"
author: team:aim
- id: aim-deck
resource: ../../aim/aim.pptx
title: "Build a Learning Guide — aim challenge pitch"
author: team:aim
---
# At a glance
| Field | Value |
|---|---|
| Organizer | [aim](/organizers/aim.md) |
| Target groups | Children and young people, plus parents looking for offers for their children |
| Focus institutions | aim, experimenta, Haus der Familie, IPAI |
| Focus topics | AI, coding, and gaming |
| Form | Open: web app, website, or another prototype that fits the target group |
| Core requirement | A child-friendly, appealing, searchable guide that is easy to maintain and easy to extend |
# The problem
Heilbronn has many institutions offering leisure activities for children and young people, from creative workshops to sports courses. Each institution publishes on its own website. Families therefore have to know which institutions exist, search several sites, and then hope that an offer matches the child's interests and is high quality.[^aim-readme]
The burden falls hardest on younger children, especially primary-school children, who cannot research on their own and depend on their parents. A poorly matched course can reduce motivation, and a child who loses the desire to try new things may miss exactly what would help them grow.[^aim-readme]
# Your task
Build the prototype of a digital application that **consolidates and presents learning and leisure offers for families in Heilbronn**. The form is up to you; it does not have to be a web app or website. The prototype should give every family member an entry point and should make COSMOS HN more accessible to everyone in Heilbronn.[^aim-readme]
The README suggests, but does not require, a first choice such as "Who are you looking for? — For me / For my child", after which the interface adapts to the group.[^aim-readme]
# Requirements
| Requirement | What it means |
|---|---|
| Target audience | Both child-friendly and appealing to adults. Children and young people want to take part; parents search on their behalf. |
| Data presentation | Not a basic list. Present offers in an appealing, innovative, creative, and target-group-appropriate way, beyond a classic event calendar. |
| Spatial focus | Limit the prototype to aim, experimenta, Haus der Familie, and IPAI. |
| Thematic focus | Limit offers primarily to AI, coding, and gaming. |
| Maintenance | Keep effort low. Ideally a person without a technical or IT background can maintain the application. |
| Scalability | Make it easy to add institutions and topics later without rebuilding the tool. |
| Functionality | Provide a good overview and a search function. Additional features are welcome. |
# Work packages
The README frames five work packages. They are suggestions, not a mandatory architecture.
## A — Surface design
Design an interface that works for children, young people, and adults. Questions from the brief:[^aim-readme]
- What does an interface for children and young people need? Should it differentiate further by age?
- What functionality lets children use the app mostly independently?
- Can one interface serve children, young people, and adults, or is a child-friendly design also appealing to adults?
- What privacy requirements affect the interface design?
Concrete guidance:
- Children read and write well only from around age 8. Use simple words, short sentences, and a larger sans-serif font. Consider letting children compose sentences from images or word choices.
- Choose metaphors and icons that make sense to the target group.
- Underage users should not be able to register themselves for events. A feature that lets a child show an offer to a parent, who then registers the child, may help.
- Consider digital accessibility throughout.
- The brief points to research on child-friendly websites, the UK children's code, the GDPR, the Digital Services Act, and the EU AI Act, and the OECD recommendation on children in the digital environment.
## B — Data input and maintenance
Show how the application receives data and stays current. The brief asks for a demonstration of automated or semi-automated maintenance, for example importing offers directly from institutions' websites.[^aim-readme]
- Check for APIs, RSS feeds, iCalendar feeds, or CSV/JSON exports before scraping.
- Decide which data is meaningful to extract from a website.
- Identify where human oversight is still necessary.
- Make it easy to add a new institution or topic.
- Check whether extracting data from each website is permitted.
## C — Data presentation
Present the collected data in a meaningful and appealing way. The brief asks for an overview that may be more innovative than a list or a classic event calendar, and for decisions about which information about an offer should be shown and how each data type should be displayed.[^aim-readme]
## D — Search functionality
Design search for the target group. The platform should include multiple themes, and users should be able to search for them.[^aim-readme]
- Make search usable for children.
- Let children and young people discover offers playfully, for example through a vision board, even when they do not know what they are looking for.
- Support filtering by themes (AI, coding, gaming), age groups, dates, and event types such as courses, workshops, and events.
## E — Data privacy for children and young people
Privacy is part of the prototype, not an afterthought. The brief asks what data is truly necessary and how to design an attractive, motivating interface that ideally stores no data at all.[^aim-readme]
- **Data minimisation:** Can search and filtering work without a user account?
- **Age differences:** Under the GDPR in Germany, young people can consent to data processing from age 16; younger users need parental consent. Consider what this means for the product.
- **Registration:** Guest or incognito mode is ideal for children, with favourites and search history disappearing when the session ends. A parent account with child sub-profiles may be useful for parents. Consider registration without an email address, using fantasy names and picture passwords such as animal symbols.
- **Consent and comprehensibility:** Explain privacy in child-friendly language and make clear when parents must be involved.
- **Location and interests:** Find suitable offers without storing exact home addresses or permanent interest profiles.
- **Gamification:** Use playful elements without collecting excessive data. Avoid leaderboards with real names and rewards for daily logins.
- **AI and chatbots:** Make clear that users are interacting with an AI, keep responses age-appropriate, check provider age limits, and do not require children to share personal details in chat.
- **External services:** Minimise data shared with third parties and consider scanning children's inputs for names, ages, or addresses before sending a request to an AI provider.
- **Security and deletion:** Define who can access data, how it is protected, and when it is deleted.
# Bonus
The README lists these as optional, not required:[^aim-readme]
- **Notifications:** Tell users when something they searched for is not currently offered, or collect anonymous search queries and wishes in a backend overview so institutions can see what is missing.
- **Guidance:** An AI assistant that lets users describe what they are looking for in their own words and suggests suitable offers, asking follow-up questions when needed.
- **Reach and accessibility:** Multilingual content, simple language, mobile-friendly design, no registration requirement, a share function for messengers, and good search-engine visibility.
- **Distance calculation:** Show how far an offer is from home or school, for example with OpenStreetMap, and help institutions see the spatial distribution of offers.
- **Expanded target group:** Include parents and educational professionals in day-care, after-school care, or schools.
# What makes a strong solution
The README says a strong solution has transparent code and fulfils all requirements, including scalability. Automated or semi-automated data maintenance is particularly compelling. It also prioritises design that fits the target audience and presents offers in an appealing way, beyond a traditional event calendar.[^aim-readme]
# Data sources
The README names these institutions and links. Teams should check each site's current structure and terms before automating collection.
| Institution | Website | Relevant page |
|---|---|---|
| aim — Akademie für Innovative Bildung und Management | https://www.aim-akademie.org/ | Holiday and leisure learning, digital education, Juniorakademie, and coding subpages |
| experimenta | https://www.experimenta.science/ | https://www.experimenta.science/lernen/ |
| Haus der Familie | https://hdf-hn.de/ | https://hdf-hn.de/kinder-jugendliche/ |
| IPAI Foundation | https://ipai-foundation.ai/ | https://ipai-foundation.ai/#events |
# Sources
The README is the detailed brief.[^aim-readme] The pitch deck states the same challenge in short form.[^aim-deck]
[^aim-readme]: aim, "Learning Guide — README", challenge brief, 2026.
[^aim-deck]: aim, "Build a Learning Guide", challenge pitch deck, 2026.
@@ -0,0 +1,110 @@
---
type: Hackathon Challenge
title: "LoRaWAN network planning: from real-world signals to smarter deployment"
description: Use real urban LoRaWAN reception data to explain signal quality, predict coverage at new locations, and build a first gateway-planning tool for cities.
resource: ../../HNVG/readme%20LoRaWAN_Network_Planning_Challenge.docx
tags: [hackathon, future-city-hackathon-2026, hnvg, lorawan, network-planning, geospatial, prediction, smart-city]
status: stable
generated: { by: opencode/deepseek-v4.1-flash, at: 2026-11-13T17:13:56Z }
sources:
- id: hnvg-readme
resource: ../../HNVG/readme%20LoRaWAN_Network_Planning_Challenge.docx
title: "LoRaWAN Network Planning Challenge — README"
author: team:hnvg
- id: hnvg-deck
resource: ../../HNVG/Pitch%20%20LoRaWAN_Network_Planning_Challenge.pptx
title: "Smart Cities Need Connectivity — challenge pitch"
author: team:hnvg
---
# At a glance
| Field | Value |
|---|---|
| Organizer | [HNVG](/organizers/hnvg.md) |
| Core question | How can a city make a better first decision about its future LoRaWAN network? |
| Data | About 600 transmitters, 8 gateways, and real-world reception data from a Heilbronn test area |
| Tasks | Understand, explore, predict, and plan |
| Output | A coverage prediction and a prototype that compares candidate gateway locations |
| Scope | Partial solutions are explicitly welcome |
# The problem
Smart meters, environmental sensors, mobility, and energy applications all need reliable connectivity. Cities considering a LoRaWAN network face practical questions: How many gateways are needed? Where should they go? What coverage can be expected?[^hnvg-readme]
Radio coverage is not a perfect circle around a gateway. Buildings, terrain, height, vegetation, and other characteristics of the urban environment influence whether a signal actually arrives. The challenge is to plan deployments from expected reception quality rather than from a simple radius.[^hnvg-readme]
# Your task
The brief defines four connected tasks. Teams may go deep in one or more of them.
## 1 — Understand
Explore the reception data, identify patterns, and develop hypotheses about why some signals are received reliably while others are not.[^hnvg-readme]
## 2 — Explore
Find and integrate publicly available data that helps explain the patterns. Possible factors include distance and obstacles such as buildings and terrain, radio settings such as spreading factor, bandwidth and transmit power, and interference or network traffic. The brief presents these as examples, not a checklist: find the data you think matters and test whether it actually does.[^hnvg-readme]
## 3 — Predict
Use what you learned to develop a model or another approach that estimates where LoRaWAN reception is likely to work and where it may become difficult.[^hnvg-readme]
## 4 — Plan
Turn the insights into a first planning tool for cities. The prototype should let users select and explore different potential gateway locations and assess expected coverage and likely gaps.[^hnvg-readme]
# The vision
The intended flow is **select an area → explore relevant data → assess potential sites → estimate reception**. The goal is not to replace professional radio planning or field measurements. It is to give a city a useful first answer to "What could a LoRaWAN network look like in our city?" and to make the approach transferable to other cities and regions.[^hnvg-readme]
# What you get
The challenge package includes six tables. The full schema, row counts, and data-quality notes are in the [HNVG LoRaWAN reception dataset](/data/hnvg-lorawan-reception.md) concept.
| File | What it contains |
|---|---|
| `transmission_data.csv` | Received uplink messages from water meters, with gateway reception, signal quality, radio settings, and message metadata. |
| `transmitter meta data.csv` | One row per water meter: meter size, installation date, calibration year, address, and GPS coordinates. |
| `gateway meta data.csv` | One row per gateway: gateway identifier, display name, latitude, and longitude. |
| `mapping transmitter id.csv` | Links a device's system ID and name to its water meter number. |
| `HN_Adresses.csv` | Addresses in Heilbronn with postal code, city, street, and house number, usable for out-of-sample predictions after geocoding. |
| `potential_gw_location.csv` | Candidate gateway locations with latitude and longitude, usable to compare sites and estimate coverage and redundancy. |
## Data caveats worth knowing before modelling
- `transmission_data.csv` is semicolon-delimited and uses a period as the decimal separator. Some metadata files use a comma as the decimal separator for coordinates, so normalise before joining.
- One metadata file is not UTF-8; the supplied copy decodes as Latin-1. Import identifiers as text to preserve leading zeros.
- Received messages alone do not establish packet-loss rates. A reliable estimate of expected transmissions is needed for that.
- Preserve separate gateway receptions of the same message when analysing redundancy; do not deduplicate them away.[^hnvg-readme]
- The README says about 600 transmitters and 8 gateways. The delivered `transmission_data.csv` contains 448 distinct devices and 9 distinct gateway IDs, while the gateway metadata has 8 rows. Check the mismatch before treating either count as ground truth.
- GPS coordinates in the transmitter metadata are of uncertain quality; the README presumes WGS 84.
# Expected deliverables
Depending on scope, the brief asks for:[^hnvg-readme]
- A working prototype, analysis, or demonstrator.
- Source code and basic setup instructions.
- A short explanation of the approach.
- An overview of additional datasets and sources.
- Key findings or results.
- Assumptions and limitations.
- A short final pitch or demo.
# What makes a strong solution
- **Smart use of data:** identify useful additional data and show why it matters.
- **Evidence:** demonstrate that assumptions or the model help explain or predict reception.
- **Practical value:** help a city make a better first deployment decision.
- **Usability:** make the result understandable without radio-planning expertise.
- **Transparency:** make assumptions, limitations, and uncertainty visible.[^hnvg-readme]
Partial solutions are welcome. Depth in one part is better than a superficial solution to everything: discovering a valuable public dataset, explaining differences in reception, building a strong coverage model, creating an intuitive visualisation, or building the full planning prototype are all valid contributions.[^hnvg-readme]
# Sources
The README is the detailed brief and data dictionary.[^hnvg-readme] The pitch deck states the challenge in short form.[^hnvg-deck]
[^hnvg-readme]: HNVG, "LoRaWAN Network Planning Challenge — README", challenge brief and data documentation, 2026.
[^hnvg-deck]: HNVG, "Smart Cities Need Connectivity", challenge pitch deck, 2026.
+8
View File
@@ -0,0 +1,8 @@
# Challenges
* [Measure where it matters](ai-traqc-sensor-placement.md) - Place stationary heat sensors and plan mobile measurement routes so a coarse satellite map becomes a meaningful fine-grained heat map.
* [Learning Guide](aim-learning-guide.md) - Build a child-friendly, low-maintenance digital guide that consolidates Heilbronn's learning and leisure offers for families.
* [LoRaWAN network planning](hnvg-lorawan-network-planning.md) - Predict LoRaWAN reception from real urban data and turn it into a first gateway-planning tool for cities.
* [WasteRouting.AI](prezero-waste-routing.md) - Turn a year of glass-container fill levels into pattern insights and a configurable, fill-level-driven route planner.
* [Pulse of the City](schwarz-digits-pulse-of-the-city.md) - Turn privacy-preserving footfall counts into a production-ready analytics and prediction app on STACKIT.
* [Riding the Sentiment](stadt-heilbronn-cable-car-sentiment.md) - Analyze two years of public debate around Heilbronn's planned cable car and make it explorable in a dashboard.
@@ -0,0 +1,113 @@
---
type: Hackathon Challenge
title: "WasteRouting.AI: fill-level-driven collection routes"
description: Analyze more than a year of glass-container fill levels, find the patterns behind them, and build a configurable route planner that recommends optimized collection runs.
resource: ../../PreZero/20261113_Readme%20Challenge%20PZ.docx
tags: [hackathon, future-city-hackathon-2026, prezero, waste-management, routing, optimization, time-series, logistics]
status: stable
generated: { by: opencode/deepseek-v4.1-flash, at: 2026-11-13T17:13:56Z }
sources:
- id: prezero-readme
resource: ../../PreZero/20261113_Readme%20Challenge%20PZ.docx
title: "WasteRouting.AI — PreZero challenge README"
author: team:prezero
- id: prezero-deck
resource: ../../PreZero/PreZero.pptx
title: "WasteRouting.AI — challenge pitch"
author: team:prezero
---
# At a glance
| Field | Value |
|---|---|
| Organizer | [PreZero](/organizers/prezero.md) |
| Data | More than a year of fill-level readings for glass recycling containers |
| Core problem | Fixed area-by-area schedules ignore how full each container actually is |
| Output | Data ingestion, pattern visualisation, and an optimized route engine |
| Key idea | Make the routing model configurable through general operational variables |
| Scope | A strong prototype concept is enough; refinement is expected after the event |
# The problem
The City of Heilbronn, like many municipalities, empties public waste bins area by area on a fixed schedule, regardless of how full each container actually is. In many cases fill-level data is neither collected nor tracked. Municipal services and private waste companies want AI route optimization, but those algorithms cannot deliver true efficiency without real-time fill-level intelligence.[^prezero-readme]
PreZero's in-house solution Fillinq can collect fill levels in a pragmatic, user-friendly way. The challenge starts where that data becomes available: what do the fill levels mean, and how should they change collection routes?[^prezero-readme]
# Your task
## 1 — Analyze
Explore the fill-level dataset and uncover key insights:[^prezero-readme]
- **Patterns and cycles:** What distinct accumulation patterns or cycles emerge over time?
- **Container grouping:** Do specific containers share similar behaviour or fill rates?
- **Anomalies and external factors:** How do school vacations, public holidays such as Christmas, Easter and Pentecost, and seasonal shifts affect waste volume?
- **Drivers:** Look beyond surface trends to understand what causes the observed patterns.
## 2 — Develop
Turn the insights into actionable intelligence by engineering a tool that serves as the foundation for dynamic, optimized route planning.[^prezero-readme]
## 3 — Define general variables
Make the solution adaptable and scalable across municipal services by exposing configurable operational variables. The brief names:[^prezero-readme]
| Variable | Why it matters |
|---|---|
| Waste stream and density | Different waste categories have different weight characteristics. |
| Container specifications | Container type and maximum volume capacity constrain how much can be collected. |
| Fleet characteristics | Vehicle types, volume capacities, and weight limits determine what a fleet can carry. |
| Inter-point distances | Travel distance between individual collection points drives route cost. |
| Disposal depot proximity | Distance from collection points to processing or unloading sites affects the route. |
| Depot distances | Distance from the central yard to the first and last stops shapes the working day. |
| Shift constraints | Maximum allowable working hours per driver or shift bound the route. |
The list ends with an ellipsis in the brief; teams are expected to add variables they find relevant.
## 4 — Route optimization
Using these inputs, the system should automatically recommend optimized collection routes. Based on calculated total weight and volume, it should also determine the optimal fleet deployment, for example whether one vehicle or multiple vehicles are required.[^prezero-readme]
# What you get
The package contains [PreZero glass fill levels](/data/prezero-glass-fill-levels.md): 11,701 readings for 449 anonymized containers, covering 2025-10-17 to 2026-09-30, with three glass colours (white, green, brown).
## Gaps to design around
The dataset is deliberately narrow. It contains only `Location_id`, `material_type`, `date`, and `fill_level_perc`.[^prezero-data]
- **No coordinates.** Container locations are anonymized IDs without latitude or longitude. Inter-point distances, depot distances, and route geometry are not in the file. A route engine needs an explicit assumption, a synthetic layout, or an external geodata source.
- **No container specifications.** Capacity and container type are not provided; they are variables the brief asks you to model.
- **No fleet or depot data.** Vehicle capacities, weight limits, shift rules, and depot locations must be configured or assumed.
- **Irregular readings.** Fill levels are recorded at irregular intervals, not as a complete daily time series. `fill_level_perc` is mostly numeric from 0 to 100 but also contains the string `overfull` (1,192 rows), which needs an explicit encoding.
- **Anonymized locations.** The IDs are stable within the dataset, so grouping and per-container histories still work, but they cannot be joined to a map without additional information.
# Expected deliverables
The brief lists four deliverables:[^prezero-readme]
1. **Data ingestion tool:** a functional interface to upload and process fill-level datasets.
2. **Pattern visualisation:** dashboards illustrating trends, anomalies, and insights.
3. **Route engine:** a working prototype that generates optimized route proposals.
4. **Documentation:** a brief user guide covering the technical architecture and approach.
# What makes a strong solution
The brief is explicit that the main goal is to see creative problem-solving and a prototype concept, not a finished product. Strong concepts could become the foundation for future development or even a startup project.[^prezero-readme]
A strong entry is therefore likely to:
- Show real insight from the data rather than only a dashboard shell.
- Make every operational assumption configurable instead of hard-coding one city's fleet.
- Handle the missing spatial and fleet data transparently.
- Connect pattern findings to routing decisions, not treat analysis and optimization as unrelated demos.
- Explain why the proposed routes are better than a fixed area-by-area schedule.
# Sources
The README is the detailed brief.[^prezero-readme] The pitch deck states the same challenge in short form.[^prezero-deck] Dataset facts come from the delivered CSV.[^prezero-data]
[^prezero-readme]: PreZero, "The Challenge — WasteRouting.AI", challenge README, 2026.
[^prezero-deck]: PreZero, "WasteRouting.AI", challenge pitch deck, 2026.
[^prezero-data]: PreZero, `glass_fill_levels_anonymized_f.csv`, delivered dataset, 2026.
@@ -0,0 +1,88 @@
---
type: Hackathon Challenge
title: "Pulse of the City: turning footfall data into decisions"
description: Build a production-ready application on STACKIT that turns privacy-preserving pedestrian counts from Heilbronn into an interactive tool for planning, and optionally into predictions.
resource: ../../Schwarz%20Digits/FCH%20Challenge%20The%20_Pulse%20of%20the%20City_.pdf
tags: [hackathon, future-city-hackathon-2026, schwarz-digits, footfall, pax-counter, smart-city, visualization, prediction, stackit]
status: stable
generated: { by: opencode/deepseek-v4.1-flash, at: 2026-11-13T17:13:56Z }
sources:
- id: pulse-brief
resource: ../../Schwarz%20Digits/FCH%20Challenge%20The%20_Pulse%20of%20the%20City_.pdf
title: "The Pulse of the City — challenge brief"
author: team:schwarz-digits
- id: digits-deck
resource: ../../Schwarz%20Digits/Challange%20Schwarz%20Digits.pdf
title: "Pulse of a City — Schwarz Digits challenge deck"
author: team:schwarz-digits
---
# At a glance
| Field | Value |
|---|---|
| Organizer | [Schwarz Digits](/organizers/schwarz-digits.md) |
| Data | Privacy-preserving PAX counter footfall data from sensors in Heilbronn's inner city |
| Output | A production-ready application, already deployed to the provided STACKIT infrastructure |
| Bonus | Vue and Go, plus the Onyx component library |
| Users | Retail store managers, waste planners, and the City of Heilbronn |
| Stretch goal | Estimations and predictions, not only historical visualisation |
# The problem
Cities and retailers want to know how many people walk down a street, but raw sensor counts are just numbers. Store managers and city planners struggle to visualise those counts alongside contextual factors such as weather or holidays. Without an intuitive way to analyse past crowds, opportunities for efficiency are lost.[^pulse-brief]
The true value is not only crunching numbers. It is proving that a team can take raw, privacy-preserving sensor data, fuse it with real-world context, and turn it into an intuitive visual story. A predictive model is the extra step that turns hindsight into foresight.[^pulse-brief]
# The data
The core dataset is provided by **Code for Heilbronn** and captured through a network of LoRa-based PAX counters deployed across the city. The sensors estimate crowd density by passively sniffing Bluetooth signals (MAC address advertisements) from nearby smartphones and smartwatches, then transmit aggregated device counts over a low-power LoRaWAN network in near real time.[^pulse-brief]
Important properties:
- The counts are **estimates**, not a census. Not everyone carries a device with Bluetooth enabled, so normalisation and critical interpretation matter.
- The sensors are described as **privacy-preserving** because they count aggregated devices rather than identify people.
- The brief advises the organizer to provide the data as clean CSV or JSON with clear geographic coordinates for each sensor. The dataset is **not stored in this repository**; teams receive it through the organizer.
# Your task
Build an interactive tool that helps different parties get insights from the data for planning, reviewing, and optimizing events, stores, and maintenance. Moving from plain numbers to estimations and predictions is explicitly called out as extra cool.[^pulse-brief]
The Schwarz Digits deck adds the delivery requirement: a **production-ready application that is already deployed to the provided STACKIT infrastructure**. Access to the STACKIT Kubernetes Engine is available by talking to the organizer.[^digits-deck]
Bonus points:
- Use **Vue** and **Go**, the stack Schwarz Digits uses for most modern applications.
- Use the **Onyx** component library at https://onyx.schwarz/.
# Who uses it
The brief describes three concrete use cases. They are useful as design targets, not as a required feature list.
| User | Use case |
|---|---|
| Lidl / Kaufland | Store managers visually identify when and why foot traffic spiked, then optimize staffing schedules. With predictions, they can adjust baking times for fresh bread or push localized app discounts. |
| PreZero | Planners analyse historical heatmaps to redraw waste collection routes around proven high-traffic zones. With predictions, they can move to dynamic routing, for example scheduling a targeted Sunday morning collection before bins overflow after a forecast sunny Saturday at the Neckarufer. |
| City of Heilbronn | Integrate the tool into a public Smart City Dashboard so citizens can see how busy areas typically are. Predictions can help manage dynamic street lighting, route traffic during major events, or place temporary infrastructure. |
# What makes a strong solution
- **Fuse data with context.** Combine footfall counts with weather, holidays, events, or other real-world signals that explain the patterns.
- **Tell a visual story.** The result should be intuitive, not a table of aggregated numbers.
- **Make it usable.** A store manager or city planner should be able to move from the overall picture to a specific street, time, or event.
- **Ship it.** The application should run on STACKIT, not only on a laptop.
- **Respect privacy.** Work with aggregated counts and avoid exposing personal information.
- **Go beyond hindsight.** A predictive or estimation layer is the differentiator.
# Open questions
- What spatial granularity and time range does the PAX dataset cover? The brief describes the source but not the exact schema.
- How should device penetration bias be corrected or at least disclosed?
- What does "production ready" mean for authentication, operations, and data refresh in the time available?
# Sources
The challenge brief contains the data description, use cases, and advice.[^pulse-brief] The Schwarz Digits deck contains the STACKIT deployment requirement and the technology bonus.[^digits-deck]
[^pulse-brief]: Schwarz Digits / Code for Heilbronn, "The Pulse of the City", challenge brief, 2026.
[^digits-deck]: Schwarz Digits, "Pulse of a City — Challenge Future City Hackathon 2026", challenge deck, 2026.
@@ -0,0 +1,171 @@
---
type: Hackathon Challenge
title: "Riding the Sentiment: public opinion on urban cable cars"
description: Analyze two years of social media and news coverage to show how Heilbronn's planned cable car is discussed, how the debate changes over time, and how it compares with other projects.
resource: ../../Stadt%20Heilbronn/readme%20Seilbahn_Challenge.docx
tags: [hackathon, future-city-hackathon-2026, stadt-heilbronn, cable-car, sentiment, stance-detection, nlp, dashboard]
status: stable
generated: { by: opencode/deepseek-v4.1-flash, at: 2026-11-13T17:13:56Z }
sources:
- id: seilbahn-readme
resource: ../../Stadt%20Heilbronn/readme%20Seilbahn_Challenge.docx
title: "Riding the Sentiment — Seilbahn challenge README"
author: team:stadt-heilbronn
- id: seilbahn-deck
resource: ../../Stadt%20Heilbronn/Pitch%20%20Seilbahn_Challenge.pptx
title: "Riding the Sentiment — challenge pitch"
author: team:stadt-heilbronn
- id: seilbahn-data
resource: ../../Stadt%20Heilbronn/FSI_Heilbronn_Medienbeitr%C3%A4ge%20(1).json
title: "FSI Heilbronn Medienbeiträge — media mention dataset"
---
# At a glance
| Field | Value |
|---|---|
| Organizer | [Stadt Heilbronn](/organizers/stadt-heilbronn.md) |
| Project | A planned urban cable car connecting Heilbronn with the IPAI campus |
| Data | About two years of public social media and news content |
| Comparison cases | Câble C1 in the Paris region, the planned cable car in Bonn, and the IPAI campus in Heilbronn |
| Output | An interactive dashboard plus three to five key insights for Heilbronn |
| Core skill | Turning thousands of individual voices into a clear, traceable picture of public debate |
# The problem
Large urban infrastructure projects trigger very different reactions. Some people see new opportunities for sustainable mobility and urban development; others question costs, benefits, feasibility, or priorities. The City of Heilbronn is planning an urban cable car as an extension of public transport, connecting the city with the IPAI campus. The question is: what does the public actually think?[^seilbahn-readme]
Raw sentiment is not enough. A negative-sounding statement does not necessarily mean opposition to the project, so the brief asks teams to consider **stance towards the project** as well as sentiment.[^seilbahn-readme]
# Your task
## 1 — Explore
Explore the public debate around Heilbronn's planned cable car. Look for patterns in social media posts, comments, and news coverage, and identify the topics, arguments, and concerns that shape the discussion.[^seilbahn-readme]
Possible questions:
- What are the main topics and discussion points?
- Which arguments are used for or against the project?
- Are reactions predominantly supportive, critical, neutral, or undecided?
- Are there recurring concerns or expectations?
- Do social media discussions differ from news coverage?
Go beyond simple positive-versus-negative sentiment and examine stance towards the project.
## 2 — Compare
Put the Heilbronn debate into perspective with other projects:[^seilbahn-readme]
- An operating urban cable car, such as the **Câble C1** in the Paris region.
- Another planned cable-car project, such as the one in **Bonn**.
- A different major development project in Heilbronn, such as the **IPAI campus**.
Questions:
- Are reactions similar across cities and projects?
- Are certain topics typical for urban cable cars regardless of location?
- Which concerns or expectations are specific to Heilbronn?
- Is an unusual transport system discussed differently from another major urban project such as IPAI?
- Do projects at different stages, from planning to construction and operation, generate different types of debate?
## 3 — Track
Public opinion is not static. Analyze how the discussion develops over time, identify moments when attention suddenly increases or reactions change, and connect those patterns with real-world developments such as:[^seilbahn-readme]
- Political decisions and announcements.
- New cost estimates.
- Route or planning changes.
- Visualisations of the project.
- Construction milestones.
- Media reports.
- The opening of a new system.
Then ask what happens afterwards: does the discussion quickly return to normal, or does the event change the debate more permanently?
## 4 — Visualize
Turn the findings into an interactive and engaging dashboard. It should let users explore the debate rather than only display a few aggregated indicators. Make visible:[^seilbahn-readme]
- How much people are talking about a project.
- What they are talking about.
- How they react.
- How opinions and topics change over time.
- Which events trigger reactions.
- How different projects and cities compare.
Users should be able to move from an overall picture to individual topics, periods, or projects and understand what is driving the patterns they see.
# What you get
The package contains [Stadt Heilbronn media mentions](/data/stadt-heilbronn-media-mentions.md): 39,361 records published between 2024-10-19 and 2026-10-27. The data was collected from publicly accessible content without requiring a login and without circumventing technical access restrictions. Direct personal identifiers are not included.[^seilbahn-readme]
The dataset was pre-filtered with a small keyword set:
- `IPAI` for discussions related to the IPAI project in Heilbronn.
- `Seilbahn` for German-language discussions related to cable cars.
- `téléphérique` for discussions around the cable car project in Paris.
Keyword filtering does not mean every post is relevant. Broad terms such as `Seilbahn` or `téléphérique` also capture tourist cable cars, toys, and unrelated projects, so further filtering and identifying which posts belong to the projects of interest is part of the challenge.[^seilbahn-readme]
## Data caveats
- The dataset is a media-monitoring export, not a representative survey. Conclusions should describe the observed conversation, not the whole population.
- The `sentiment` field uses a small integer scale and includes `-99` as a distinct sentinel value. Treat it as unknown or unclassified rather than as an extreme negative score unless you can verify its meaning.
- `reach` is missing for some records; `engagements` is available but may be zero or platform-specific.
- Records span many media types (news, YouTube videos and comments, forums, blogs, social networks) and several languages. Normalise before comparing.
- Duplicate or near-duplicate content is possible across sources.
# Methods
The brief says there is no single correct approach. Suggested methods include:[^seilbahn-readme]
- Natural Language Processing.
- Sentiment analysis.
- Stance detection.
- Topic modelling.
- Embedding-based clustering.
- Keyword and argument analysis.
- Named entity recognition.
- Temporal analysis.
- Peak or change-point detection.
- LLM-based text classification.
- Interactive data visualisation.
Whatever the approach, results must remain understandable and traceable. A sophisticated model is only useful if users can understand what the results mean.[^seilbahn-readme]
# Expected outcome
The brief lists these expected elements:[^seilbahn-readme]
- A reproducible approach for processing and analyzing unstructured text data.
- A meaningful classification of the major discussion topics.
- An analysis of sentiment and/or stance towards the projects.
- A temporal view of how the debate develops.
- Identification of relevant events and reaction peaks.
- Comparisons between cities and projects.
- An interactive dashboard that communicates the results clearly.
- Three to five key insights that Heilbronn can take away from the analysis.
The final result should combine data analysis and storytelling: not just numbers, but the patterns behind them.[^seilbahn-readme]
# Responsible data use
Teams may collect additional data, but the brief requires them to respect data protection and privacy, comply with platform terms and technical access restrictions, collect only what is necessary, avoid exposing personal information in the dashboard, and document where additional data comes from and how it was collected.[^seilbahn-readme]
# What makes a strong solution
- A clear, defensible definition of relevance, topic, sentiment, and stance.
- A pipeline that another person could reproduce.
- A temporal story tied to real events, not just a line chart.
- Comparisons that explain what is specific to Heilbronn.
- A dashboard that supports exploration from overview to detail.
- Insights a city administration could actually use.
# Sources
The README is the detailed brief.[^seilbahn-readme] The pitch deck summarises the challenge.[^seilbahn-deck] Dataset facts come from the delivered JSON export.[^seilbahn-data]
[^seilbahn-readme]: Stadt Heilbronn, "Riding the Sentiment: Exploring Public Opinion on Urban Cable Cars", challenge README, 2026.
[^seilbahn-deck]: Stadt Heilbronn, "Riding the Sentiment", challenge pitch deck, 2026.
[^seilbahn-data]: FSI Heilbronn, `FSI_Heilbronn_Medienbeiträge (1).json`, delivered media mention export, 2026.