SentioAir Research

Methods and standards

Research is only useful if people can check it. This page sets out how we measure, how we report and what we make available.

i.

Measure first

Claims follow data, at the resolution the question needs.

ii.

Publish the method

Every result ships with how it was produced and how to reproduce it.

iii.

State the limits

Uncertainty and limitations are reported as plainly as findings.

iv.

Validate

Low-cost measurements are checked against reference instruments.

The air quality index we use

Every index we report uses the Indian National Air Quality Index from the Central Pollution Control Board (CPCB). Band names are used exactly as CPCB defines them. The table shows the PM2.5 breakpoints, which apply to a 24-hour average.

BandPM2.5, 24-hour (µg/m³)Sub-index
Good0 to 300 to 50
Satisfactory31 to 6051 to 100
Moderate61 to 90101 to 200
Poor91 to 120201 to 300
Very poor121 to 250301 to 400
SevereAbove 250401 to 500

Where we show a band for a shorter period, such as a single reading, we say so, because the official index is defined on 24-hour averages.

Reporting uncertainty

Each paper states its limitations as plainly as its findings: the instrument used, how long it ran, how much data was captured and what the measurement cannot show.

Validation

Low-cost measurements are checked against reference instruments. A co-location study comparing SentioAir monitors with reference instruments is in preparation and will be published here.

Data and code

Where we can, the data behind a paper is listed on the datasets page under an open licence, so that others can reproduce the result.