In index we trust

PSA: the NEA has reworked its PSI readings page again. It is now deeper in the site.

So obviously there is a lot of discussion about the Pollutant Standard Index (PSI, not PSI Index please). It is, as some have quipped, an index even more closely watched than some of the the stock markets around the world. But is it really relevant?

First, Minister for Environment and Water Resources (EWR) Dr. Vivian Balakrishnan told us not to be too fixated on the PSI data. Never mind the fact that we are hardwired to look at results “since young” (sic). Then, Second Minister for EWR, Ms. Grace Fu told us we were preoccupied with the wrong data – that the 3-hour average is less important than the 24-hour one. Yes, the 24-hour average is how much we’ve been going through over the past day, but isn’t the more current information directly relevant to deciding what you do next?

So the EWR ministers are saying “no, it is not relevant”, and “well, maybe it is relevant but you should be looking at this instead”. Which is it, then? Well, as it turns out, the simple metric of “ohmygod the air is on fire” or “is someone grilling something in here” is probably the most reliable. Sometimes, it is best to rely on gut feeling and play it safe.

What is being measured anyway?

The pollutants measured in the PSI are particles up to 10 micrometres (µm) in diameter (PM10) which given the amount of trouble we are in right now we can directly see, most of the time, but there seems to be more danger present (and clear, as it were) recorded in the PM2.5 count.

The PM2.5 count is the mass of particles up to 2.5 µm in diameter in a given cubic metre of air (µg/m³) and those are fine enough to enter directly into our bloodstream via the alveoli (OK, the lungs) and are thus more dangerous. This count is separate from the PM10 reading and is updated hourly in an RSS feed (though you will only get three data points in this second table here).

One reason why I will continue keeping track is this open attempt to follow and deconstruct the PSI data given by the NEA. You might see some anomalies there, and my friend Shen Ting has pointed out why a moving average is a problem. He also shows what seems to be a better model. His friend Gerald has gone one better to find a weighted average model that extrapolates the (formerly) missing data points. You’ll have to ask him to explain that, though.

Math geekery aside, I have to be fair to the NEA and commend them on quickly providing the formerly missing data for 2am – 5am. Of course, the question would then be why were they not there before? …but let’s keep that for later, when the smog clears.

Most important thing right now, guys, is to take care of yourselves. Stay safe 🙂

This entry was posted in Just Me, Singapore. Bookmark the permalink.

2 Responses to In index we trust

  1. Shen Ting says:

    You seem to have missed out my disclaimer that I don’t think either model is actually the “right” one 🙂

    • icedwater says:

      Yes, of course, but I wrote “seems to be” and I thought that was enough to cover it.

      I was hoping people to go over and read your blog too, or I wouldn’t have bothered linking it.

      But yeah, I should probably have mentioned it all the same.

Leave a Reply

Your email address will not be published. Required fields are marked *