Yet Another Haze Index (Yahi)

<table style=”border: hidden;” bgcolor=#DDDDDD>

</table>

Overview

Yahi is an experiment in using very-low-cost dust sensors to collect and map more detailed information about Singapore’s haze than can cost-effectively be collected with lab-quality instruments:

• If it’s successful, it will make it possible to see where the haze is and how its moving and therefore expand the usefulness of the NEA’s published data in making decisions about outdoor activity during hazy periods.
• If it’s not successful, it will at least produce maps of Singapore covered in coloured circles.

I am looking for volunteers to purchase a sensing station and install it on a balcony or a roof to provide data for the project. The sensing station is available from 12Geeks as an assembled and tested kit for $95 (plus a plug pack if you don’t have a spare one lying around).

Objectives

This project has two objectives during Singapore’s haze season:

• To measure haze intensity on a minute-by-minute basis to help short-term decision making about strenuous outdoor activity, for example “should I go for a run now?”.
• To visualise the propogation of haze across Singapore.

Approach

The basic approach is to:

• establish a network of very low cost dust sensors spread over much of Singapore’s landmass,
• collect and process samples from these sensors on a minute-by-minute basis, and
• process that data to produce and share useful information about the present state of the haze.

These low cost sensors lack the precision of lab instruments but are expected to provide a sufficiently accurate proxy to meet the project’s objectives.

Status

The initial rollout of sensors is underway:

• Assembled, tested kits are available from 12Geeks.
• An approximate mapping between dust sensor outputs and NEA’s published data has been constructed.
• Several sensors have already been deployed and are feeding near-real-time data (30-90s delay) to the map above.

Next steps include:

• Complete the rollout.
• Create some more useful visualisations.

Getting Involved

If you’d like to host a sensor, particularly if you’d like to do so in a part of the map that currently has no coverage, please see the mapping pilot page.

Details

Recent State of Fires in Region

TODO: Put regional extract from NASA’s Global Fire Maps here.

Sensing Station

The current components are:

• A low-cost sensor: Shinyei PPD42NS/NJ. $21
• A WiFi-enabled micro-controller: Spark Core. $55
• A simple plastic case and some cable ties. ~$5

</img>

They are available as an assembled, tested kit from 12Geeks. $95

If you don’t have one lying around, you’ll also need a USB plug pack. $9

Depending upon how you’re installing it, you might prefer a more compact USB plug pack like the one in the photo above. $20

Data Uploading

The current plan is:

• A sketch to run the Shinyei sensor derived from Chris Nafis’ approach delivering a sample every 30s.
• The use of Spark Labs’ publish mechanism to deliver samples from the sensors every 30s. If you register an account, you can already monitor the raw sample stream at https://api.spark.io/v1/devices/events/yahi/?access_token={your_access_token} (more)
• A script running somewhere to collect the above stream and package it for delivery to the web.

Data visualisation

For last-24-hours haze intensity trend graphs:

• The intended webpage tool is D3 drawing on the aggregate data directly.
• If there is sufficient interest and volunteer app developers step forward then mobile apps drawing on the same data are also possible.

For propogation visualisation, the initial approach is to use Google Maps javascript API to draw circles directly on the map of Singapore as at the top of this page, and then display a last-24-hours animation.

A map snapshot for the benefit of web-crawlers that can’t “see” Google Maps: