Crunching large data sets to identify air pollution hot spots
Interested in using data to solve healthcare challenges?
We all have heard that air quality is becoming worse and lot of us dismiss it that it’s happening somewhere else or its a Delhi problem only! However, the reality is that bad air quality is much closer home than we want it to be.
To find out how bad it is, we need to have a good data about air quality. It has been estimated by researchers that Bengaluru needs to have at least 41 real-time air quality monitors (after taking into consideration WHO & CPCB guidelines, land use pattern, population density, air shed etc).
This is still much lower than comparable cities like London and Paris which have 120 and 60 real-time monitors respectively. Currently Bengaluru has only 10 such real-time monitors! (I’m discounting the manual stations in BLR)
Even if we have all the 41 fixed stations, to find air quality at any given point in time, we need high quality models which take in data from such stations & other parameters and estimate the air quality at any point in between these stations.
Urban air pollution concentrations vary sharply over short distances (≪1 km) owing to unevenly distributed emission sources, dilution, and physicochemical transformations. Accordingly, even where present, conventional fixed-site pollution monitoring methods lack the spatial resolution needed to characterize heterogeneous human exposures and localized pollution hotspots. This is accentuated in an Indian context where land use patterns are not well known & adherence to zoning regulations are absent in many cases.
To help estimate air quality in hyper local level & identify hot spots researchers in Oakland, California stacked a Google Street View car with good quality air quality sensors & drove them around, more details of this fascinating work is here. The way shown by the researchers in CA may be the way for us to solve the data availability problem in India and derive insights which will push us to take air quality related issues more seriously.
It is being hoped that this high resolution mobile monitoring along with large number of static monitors being facilitated as mentioned above will help providing a very accurate & personalised information about air pollution to citizens.
Clean Air Platform-Bengaluru is planning to run a pilot in Namma Bengaluru on the lines of the work done in CA. It will be running a vehicle similar to this, since we don’t have Google Street View cars 🙁
The challenge for all of us in BLR is to crunch the data generated & identify the hostspots, create a model to estimate air quality at given point in the city, and of course use all this to bring about a change, which will improve air quality!
Even before the pilot starts, the vehicle with equipment will be driven around BLR for 200 km every day (is that possible?). To start off some questions which need to be answered are:
- What is the route on which the vehicle should run? (so that max number of schools, business parks / districts & residential areas can be covered)
- How many shifts should the car be run? (we all know the legendary traffic in BLR)
3. What is the daily schedule for the car to cover a large area
4. How should we cover an area? Go over & over the same place on consecutive days or spread over a period of time?
Lets hack our way to the answers!
This Hacknight is hosted by HasGeek in association with WalmartLabs. The Fifth Elephant is India’s best big data and machine learning conference. It is a conference for practitioners by practitioners., register now for The Fifth Elephant 2018 now!
