Skip to content

Participation & Constraints

Binding guardrails

  • Only publicly available data sources and documents may be used.
  • Internal, operational, sensitive, or classified data are not part of this challenge.
  • Every source must be documented with URL, format, access path, and retrieval date.
  • Official APIs or official downloads should be clearly preferred over scraping.
  • LLM-based outputs are only acceptable when the underlying sources remain visible and traceable.
  • The focus is a small, working MVP.
  • Code, schema, access logic, and documentation should be reusable after the hackathon whenever possible.

Mandatory for all teams

Participants must source and document their own public data sources. This website provides orientation and examples, but not a complete dataset.

Expected way of working

Topic Expectation
Data selection deliberately limited, analytically justified source base
Data integration traceable mapping and transformation steps
Transparency visible evidence and clear source references
Technology pragmatic, maintainable prototype instead of over-engineering
Documentation compact but reproducible

Evaluation focus

The exact jury setup may vary, but the following points should be visible in a strong submission for this challenge:

Focus area What matters
Problem fit clear alignment with prospective environmental analysis and resilience indicators
Source clarity clean documentation, traceable origin, and access paths
Integration quality robust data model and consistent processing across multiple sources
Working MVP demonstrable end-to-end flow with a usable interface
Analytical value understandable indicators, trend cards, or weak signals with evidence
Reusability architecture, code, and documentation remain reusable
Visualization information is clear and explainable for professional users

Practical recommendation

  • limit the scope early
  • stabilize interfaces and the data model first
  • carry source references from the start
  • only scale visualization once the evidence chain is reliable

This increases the chance that the result is not only demonstrable, but also analytically usable afterwards.