Messy air-quality tables in. Decisions you can defend out.
CivicIQ helps ward offices, housing societies and schools turn the monitoring data they already have into validated numbers, findings traceable to source rows, and clear actions for the people they look after.
- Good0–50
- Satisfactory51–100
- Moderate101–200
- Poor201–300
- Very poor301–400
- Severe401–500
Results use the official CPCB AQI categories, and only when the data meets the method's requirements.
Code calculates every number
The language model only interprets what was computed.
Every finding links to its source rows
Open a finding and see the exact cells behind it.
Gaps are flagged, never filled in
Sparse data is marked insufficient instead of estimated.
How a table becomes a decision
Each step says who did the work: you, the code, or the AI.
Bring the data you have
Upload a CSV, Excel sheet, JSON file, PDF or a photo of a table, or paste rows straight from a spreadsheet. The original file is preserved.
Confirmed by you
Confirm what each column means
CivicIQ proposes a mapping for stations, timestamps, pollutants and units. Nothing is validated until you confirm it.
Confirmed by you
Validate data quality
Missing readings, duplicates, outliers, unknown units and bad timestamps are flagged row by row. Gaps are reported, never filled in.
Computed by code
Calculate the analytics
Daily means, exceedances of NAAQS and WHO guideline values, CPCB AQI, trends, spikes and hotspots are calculated the same way every time.
Computed by code
Read evidence-backed findings
A language model interprets the calculated metrics. Every finding cites the metrics it rests on, and you can open the source rows behind them.
Written by AI from the computed metrics
Get actions, a briefing and a citizen advisory
Recommendations for your role, an internal briefing note, and a public advisory in English, Hindi or Marathi at three reading levels, with care guidance for children, older adults and people sensitive to poor air.
Written by AI from the computed metrics
Compare with the previous report
See what improved, what worsened and which hotspots are new, with a clear verdict on whether the two periods are comparable at all.
Computed by code
What this demo is, and is not
- The bundled sample datasets are synthetic demonstration data, not actual observations, and are labelled as such everywhere they appear.
- Hotspots describe where readings are elevated, not why. CivicIQ does not attribute pollution to sources.
- When data is too sparse for a calculation, the result is marked insufficient instead of being estimated.
- AI-written text is labelled, and any figure in it that cannot be matched to a calculated metric is flagged for you to check.