🌍 Inspiration
Climate-related decisions depend on trustworthy data. Unusual emissions readings, inconsistent renewable-energy metrics, and changes to recorded values can make climate data harder to review confidently. We built ClimateProof AI to bring anomaly detection and tamper-evident record verification into one practical application.
💡 What It Does
ClimateProof AI is a full-stack prototype that helps users analyze climate datasets and review unusual observations. Users can upload CSV files, validate their data, inspect emissions and renewable-energy metrics, identify anomalies, export reports, and register records for later verification.
Its hash-linked audit trail records registered entries and related activity. The verification workflow can identify record mismatches, while a tamper demonstration shows how altered copies can be detected.
⚙️ How We Built It
The frontend uses React, TypeScript, Vite, and Tailwind CSS. The Python FastAPI backend handles CSV validation, analysis, record verification, and audit operations. Pandas and NumPy support data processing, while Scikit-learn provides Isolation Forest-based anomaly detection with robust statistical confirmation.
🧪 Testing and Results
We tested the prototype with a synthetic dataset containing 90 records from three locations. The analysis identified two deliberately unusual observations: elevated emissions at Chennai Plant and a low renewable-energy value at Cuddalore Facility. We also tested record registration, successful verification, duplicate rejection, tampering detection, and hash-chain integrity.
These results demonstrate the intended behavior on our test data; they do not establish real-world model accuracy.
🚧 Challenges and Learning
A key challenge was connecting data analysis with record verification while keeping the workflow understandable. Testing duplicate registrations, modified records, and audit-chain integrity helped us better understand data validation, anomaly detection, hashing, and traceability.
🔭 Future Scope
Future improvements include testing with trusted real-world climate datasets, expanding automated tests, improving anomaly explanations, and strengthening audit-log protection.
ClimateProof AI is a prototype for climate-data analysis and tamper-evident verification. Its local hash chain is not a public blockchain and does not independently certify climate measurements.
Built With
- artificial-intelligence
- data-analysis
- fastapi
- machine-learning
- numpy
- pandas
- python
- react
- scikit-learn
- tailwind
- tailwind-css
- typescript
- vite
