CROWD4SDG · Project

AI-Powered Citizen Science Tools to Track Climate Risks Using Social Media Data

environmentPrototypeTRL 4Thin data (2/5)

Imagine millions of people posting photos and reports about floods, heatwaves, and storms on social media — but nobody is collecting that data in a useful way. CROWD4SDG built machine learning tools that mine social media posts and crowdsourced reports to track climate disasters as they happen. Think of it like a weather radar, but powered by real people on the ground instead of satellites. The team also ran three annual challenge events where teams developed new citizen science projects focused on climate action.

By the numbers

232

SDG indicators that need measurement, many lacking data collection methods

UN Sustainable Development Goals addressed by the broader measurement challenge

169

concrete SDG targets underpinning the goals

annual challenge-based innovation events run to test citizen science approaches

total project deliverables produced

consortium partners across 4 countries

The business problem

What needed solving

Many companies in insurance, real estate, agriculture, and municipal services need real-time, ground-level data on climate events — but official sources are slow, expensive, and often lack local detail. With 232 SDG indicators needing measurement and many countries lacking data collection capacity, there is a massive gap in actionable climate impact data that affects business decisions from risk pricing to disaster response planning.

The solution

What was built

The project built machine learning tools for mining social media and non-traditional data sources for climate event monitoring, a visualization interface for team analytics and in-situ collected data, and methodologies for running citizen science challenge events. Across its 3-year run, the team produced 26 deliverables.

Audience

Who needs this

Property and casualty insurance companies needing real-time climate event dataEnvironmental consulting firms doing climate risk assessmentsMunicipal disaster response and civil protection agenciesESG reporting teams at large corporations tracking climate indicatorsInternational development organizations monitoring SDG progress

Business applications

Who can put this to work

Insurance & Reinsurance

enterprise

Target: Property and casualty insurers needing real-time climate event data

If you are an insurance company dealing with rising climate-related claims and slow damage assessment — this project developed machine learning tools that mine social media data to detect and monitor extreme climate events in real time. Instead of waiting days for official reports, you could get crowd-verified impact data within hours. The project ran 3 annual challenge-based innovation events that tested these tools across multiple climate scenarios.

Environmental Consulting

mid-size

Target: Climate risk advisory firms serving municipalities and corporations

If you are a climate consultancy struggling to gather ground-level impact data for your risk assessments — this project built AI tools that extract climate event information from non-traditional data sources like social media and citizen reports. The consortium of 7 partners across 4 countries tested these approaches specifically for SDG 13 climate action monitoring. This could supplement your existing data with real-time community-level observations.

Smart City & Municipal Services

any

Target: City governments and disaster response agencies needing community-level climate monitoring

If you are a municipal disaster response team that lacks granular, real-time data on how climate events affect neighborhoods — this project developed crowdsourcing methods and a visualization interface for tracking community-reported climate impacts. With 232 SDG indicators to measure and many countries lacking data collection capacity, these citizen science tools fill critical monitoring gaps for local governments.

Frequently asked

Quick answers

What would it cost to license or adopt these tools?

The project was funded as a Research and Innovation Action (RIA), meaning outputs are typically open-access. Specific licensing terms are not detailed in the available project data. You would need to contact the University of Geneva coordinator to discuss commercial use terms.

Can these tools work at industrial scale for a large insurance or consulting operation?

The tools were tested through 3 annual challenge-based innovation events and produced 26 deliverables including a visualization interface. However, this was a research project with no industrial partners in the consortium, so scaling to enterprise-level operations would likely require additional engineering and integration work.

What is the IP situation — can we use the technology commercially?

As an EU-funded RIA project, intellectual property typically remains with the consortium partners. The consortium is entirely academic (4 universities, 3 research organizations), so licensing discussions would go through university technology transfer offices. No commercial deployment evidence exists in the project data.

How reliable is crowdsourced climate data compared to official sources?

The project specifically researched quality assessment methods — rigorously assessing the quality of scientific knowledge produced by citizen science teams. Based on available project data, quality validation was a core research focus, but the project does not claim parity with official statistical data.

What specific technology was built — software, platform, or methodology?

The project built a visualization interface for team analytics and in-situ collected data, along with machine learning tools for social media data mining. With 26 total deliverables produced, the outputs span both software tools and methodological recommendations for national statistical offices.

Is this ready to deploy in our organization today?

Based on available project data, this is at the tested-prototype stage. The project closed in April 2023 with working tools demonstrated through challenge events, but there is no evidence of commercial deployment or pilot testing with paying customers. Further development would be needed for production use.

Consortium

Who built it

The CROWD4SDG consortium consists of 7 partners across 4 countries (Switzerland, Spain, France, Italy), with a purely academic composition: 4 universities and 3 research organizations. There are zero industrial partners and zero SMEs, which is a significant gap for commercial exploitation. The coordinator is the University of Geneva, and the consortium includes UNITAR (a UN training institute) as a direct channel to national statistical offices. This composition is strong for research credibility and policy influence but weak for market readiness — any company looking to adopt these tools would need to bridge the gap between academic prototype and commercial product themselves.

UNIVERSITE DE GENEVECoordinator · CH
AGENCIA ESTATAL CONSEJO SUPERIOR DE INVESTIGACIONES CIENTIFICASparticipant · ES
ORGANISATION EUROPEENNE POUR LA RECHERCHE NUCLEAIREparticipant · CH
UNIVERSITAT ZURICHthirdparty · CH
POLITECNICO DI MILANOparticipant · IT
UNIVERSITE PARIS CITEparticipant · FR
UNITED NATIONS INSTITUTE FOR TRAINING AND RESEARCHparticipant · CH

How to reach the team

University of Geneva, Switzerland — reach out through their technology transfer office or the project website contact page

Order a report on this consortium See UNIVERSITE DE GENEVE profile →

Next steps

Talk to the team behind this work.

Want to explore how citizen science and AI-powered climate monitoring tools could fit your risk assessment workflow? SciTransfer can connect you directly with the research team and help evaluate commercial licensing options.

Order a report on this topic More in Environment & Climate