Re: AQ-NEST: Air Quality Network Evaluation & Siting Tool
15 views
Skip to first unread message
Sarath Guttikunda
Aug 5, 2026, 9:42:50 AM (7 days ago) Aug 5
to data...@googlegroups.com
Dear All,
Good evening. We run an air quality concepts-building interactive course for all audiences (links below).
During the course, one is a mapping exercise. In class, participants are given a map with locations of major industries, commercial/road density, and wind rose patterns and use practical thumb rules to design a monitoring network. There are no wrong answers in this exercise. At the end, they present their approach: how they split the airshed and arrange their 2 or 10 or 20 points for the best representation. I clap, offer suggestions, and we have some back-n-forth. The exercise remains a hands-on way to understand both the thumb rules and the city's layout.
This AQ-NEST tool (all on github) is designed as a follow-up to this exercise, aimed at visualizing what they designed and evaluating the exposure and spatial coverage of the network. For us (researchers and policy makers), health exposure is what matters -- if monitoring data is going to drive discussions, is it giving a representative sample? Is it meeting the minimums? Is it covering a meaningful share of the population base?
With 100 locations, the possibilities are endless -- you can design a grid, straight lines, a flower pattern, circles, you name it, all will work. But with only five stations, where do you place them? The default is often your own rooftop and there's nothing wrong with that, given the safety, internet, and electricity access it offers. What we want participants to understand is whether a different design could offer better representation for the airshed from an exposure standpoint -- maybe moving two stations 500 meters to the left improves coverage by 5% or what happens if I add 5 more. This isn't an automated algorithm: in class, have a group discussion and place your points on a map, then place the same points in the tool and see what comes out. Then move around to see if the score can be improved. It's about getting a feel for the area and its activities -- manual input paired with manual learning.
You can use this to evaluate a random network design (by placing your points in a box) or upload one to evaluate an existing one. Sharing below two sets of images
This AQ-NEST tool (all on github) is designed as a follow-up to this exercise, aimed at visualizing what they designed and evaluating the exposure and spatial coverage of the network. For us (researchers and policy makers), health exposure is what matters -- if monitoring data is going to drive discussions, is it giving a representative sample? Is it meeting the minimums? Is it covering a meaningful share of the population base?
With 100 locations, the possibilities are endless -- you can design a grid, straight lines, a flower pattern, circles, you name it, all will work. But with only five stations, where do you place them? The default is often your own rooftop and there's nothing wrong with that, given the safety, internet, and electricity access it offers. What we want participants to understand is whether a different design could offer better representation for the airshed from an exposure standpoint -- maybe moving two stations 500 meters to the left improves coverage by 5% or what happens if I add 5 more. This isn't an automated algorithm: in class, have a group discussion and place your points on a map, then place the same points in the tool and see what comes out. Then move around to see if the score can be improved. It's about getting a feel for the area and its activities -- manual input paired with manual learning.
You can use this to evaluate a random network design (by placing your points in a box) or upload one to evaluate an existing one. Sharing below two sets of images
1. Pune airshed network evaluation (official stations), scored 28% population coverage. Relocating the same stations, score increased to 49%.
2. Mumbai with only central network and with central+BMC (proposed and waiting)
2. Mumbai with only central network and with central+BMC (proposed and waiting)
AQ-NEST tool
https://urbanemissions.info/tools/aqnest-network-evaluation-siting-tool/
Course page with summaries from past events
https://urbanemissions.info/blog-pieces/air-quality-information-awareness-for-better-planning/
(execution is customized from high-schoolers to public-body scientists)
Primer on air quality monitoring
https://urbanemissions.info/publications/primer-on-air-pollution-monitoring/
Available in multiple languages - hindi, marathi, punjabi, and bengali
Happy to do a 15-20 min show and tell, if there is general interest.
Shout out to Sai Dammalapati (in the group) for pulling this together.
With best wishes,
Sarath
With best wishes,
Sarath
Reply all
Reply to author
Forward
0 new messages