River AI Rangers gives secondary school students an AI-powered tool to investigate real water-quality data from local rivers — and ask the questions that matter.
A free Streamlit app that puts real Thames21 river data in students’ hands — with an AI tutor built in to help them interpret it.
Choose from preloaded Thames21 and Lea Guardians datasets, or upload your own CSV from any river monitoring programme.
The AI reads the actual dataset and answers questions about it — citing real numbers, not generic facts. Students can push back and question the AI too.
Compare sites, spot pollution patterns, write letters to the Environment Agency. The AI helps students find the right questions — not just the answers.
No setup. No accounts. No data leaving the classroom without your knowledge.
Choose a preloaded Thames21 river dataset or upload your own CSV. The app reads the columns and prepares the AI automatically.
Type a question or click a prompt card. The AI responds based on the real data in front of it — phosphates, ammonia, pH, dissolved oxygen.
Push back, dig deeper, compare sites across seasons, and use the data to take action — write a report, a letter, or a class presentation.
No subscription, no per-student licence, no hidden cost. The app is open-source and free to use for any school.
The AI is instructed to redirect off-topic questions, never discuss personal topics, and remind students it can make mistakes.
Instant visual summary of average and peak readings for pH, phosphates, ammonia, nitrates and dissolved oxygen — with healthy-range flags.
If your school has conducted its own river monitoring, upload the CSV and the AI analyses your data — up to 5 MB.
20 ready-made question starters across five categories: The Numbers, Compare Sites, Pollution, Take Action, and Question the AI.
Powered by Claude Haiku, responses arrive in seconds. A 30-question per session limit keeps API costs manageable for free use.
Runs entirely in the browser. Students just need the link. Works on school laptops, tablets, even phones.
Fork it, modify it, run your own instance with your own API key. The code, data, and lesson materials are all on GitHub.
All preloaded datasets come from real monitoring programmes run by Thames21 and the Lea Guardians — the same data that informs actual environmental decisions.
Focused dataset across three sites on the River Ching in East London, perfect for comparing upstream vs. downstream pollution patterns in a single lesson.
Multi-site, multi-season dataset from the River Lea catchment. Rich enough for extended investigations and controlled variable comparisons.
Synthetic dataset with clearly varied readings, designed for first lessons or demos where you want predictable, illustrative results.
Works as a one-off activity or a multi-week project. Ties directly to curriculum objectives across several subjects.
The app is live right now. Share the link with your class, pick a dataset, and start asking questions. No account required for students.
Open the app →Want to add your school’s river data?
Export your monitoring results as a CSV and upload it directly. The AI will read your column names automatically.
Free, open-source, and live right now. No setup. No account. Just real river science.