A haze reading is useful only when its label, place and time travel with it. A screenshot showing a large number may look urgent, but it cannot explain whether it is a concentration, an index, an observation or a forecast. This guide explains how to read Singapore’s official air-quality information, compare like with like and share a clear update. It does not display live air quality or replace the current health advisory.
Start with the measurement, not the colour
NEA’s air-quality FAQ distinguishes the one-hour PM2.5 concentration from the rolling 24-hour PSI. It reports five regions and explains that PSI uses pollutant sub-indices, with the highest determining the index. For immediate outdoor decisions, use the one-hour reading and corresponding guide; for next-day planning, use the PSI forecast and advisory. Different countries’ index systems are not automatically comparable.
That short distinction should stay beside any chart you share. A colour is a presentation choice, not a unit of measurement. If a website displays a red tile, read the metric and its legend before comparing it with another red tile elsewhere. Identical colours can label different scales, while different colours can appear on dashboards built from related data.
Think of the number as a sentence with missing words. The complete sentence needs to say what was measured, where, over what period and when. Removing those words does not make an update more accessible; it makes it easier to misunderstand. A clear caption often matters more than a large headline number.
Use the dated October update as an example
NEA’s 24 October 2026 advisory forecast PSI of 80–130 for 25 October. It also reported, at 5pm on 24 October, regional one-hour PM2.5 concentrations ranging from 33 to 97 micrograms per cubic metre and 24-hour PSI of 108–151. These are historical figures in this guide. They are not the reading for the moment you opened the page.
The forecast and observations belong to different time references. The publication date is when the advisory was issued; the observation time identifies a reported snapshot; the forecast date identifies the period being discussed. A message that drops those distinctions can accidentally present yesterday’s observation as tomorrow’s forecast or today’s live reading.
For a practical note, write those three fields separately. If you are referring to an old article for background, call it background. If you need a current decision, open the [official haze portal](https://www.haze.gov.sg/) or myENV rather than forwarding the old numbers again. An article can remain useful as an explanation after its example has ceased to be current.
Match the advisory to the person and activity
MOH’s haze page, updated 21 October 2026, links immediate outdoor activity to the one-hour guide and next-day activity to the PSI forecast. Its advice varies with health status, duration and exertion. People who feel unwell, particularly vulnerable groups, should seek medical advice. Read the full current table for the relevant group rather than treating this data-reading guide as an individual activity prescription.
An organiser should identify the actual question before opening a dashboard. Is the group making tomorrow’s plan or deciding whether to begin something now? Is this a short transfer between buildings or a prolonged outdoor activity? Those questions help locate the right official guidance; they do not license an organiser to invent a personal safety threshold for everyone attending.
Keep the source link in the event message and make clear who will review it again. A plan made yesterday can require another look before departure. Equally, a new screenshot does not automatically explain what action follows: it must be read alongside the applicable advisory and the organisation’s own procedures. The useful output is a clear, accountable update, not merely a forwarded colour code.
Read the region as carefully as the timestamp
The official PM2.5 dataset supplies readings for Singapore’s major regions, with regional metadata available through its documented API. A regional value should not be silently relabelled as a measurement inside a particular room, classroom or workplace. The name of the region tells the reader the geographic scope of the information.
Consider a fictional message from someone living in one region but attending an activity in another. The reading nearest home may not be the relevant entry for the destination. The message should identify which place the plan concerns, then point to the appropriate official regional information. This is a location-checking example, not a claim that any two regions have particular pollution levels today.
If a location is unfamiliar, consult the official region guide instead of guessing from a loosely drawn social-media map. Do not crop a screenshot so tightly that the region label disappears. A beautifully formatted number with its geography removed is less useful than a plain text entry with the correct label intact.
Refresh frequency is not the averaging period
The data.gov.sg PM2.5 catalogue describes updates every 15 minutes, while the readings are identified as one-hour PM2.5. These labels answer different questions. How often a service updates is not the same as the period represented by the measurement. A refreshed display does not become an instantaneous measurement merely because the browser has just loaded it.
Imagine a fictional dashboard that redraws every minute while receiving less frequent source updates. Its animation may look live, but the underlying observation has not necessarily changed. A reader should look for the observation timestamp, not infer freshness from a moving icon or countdown. If the page exposes both retrieval and observation times, preserve both when making a record.
This distinction also prevents unnecessary alarm about repeated values. An unchanged number on consecutive screen visits might reflect an unchanged published observation or a similar measurement. It does not, by itself, prove that the sensor has failed. Conversely, a changed number does not prove that a website is using the correct source. The labels and timestamps remain necessary either way.
Compare two screenshots without making a false trend
Before calling two numbers a rise or fall, check that they describe the same metric, unit, region and averaging period. Then place them in time order. If one is an index and the other a concentration, subtraction is not a meaningful comparison. If the regions differ, the difference is not automatically a change over time at one location.
Here is a deliberately hypothetical example. A chart lists the same region’s concentration as 40 and later 60 micrograms per cubic metre. The numerical increase is 20, or 50% relative to 40. That arithmetic does not mean a person’s health risk has increased by exactly 50%, and it does not establish the cause of the change. Interpretation still needs the applicable guidance and context.
Do not choose only the two most dramatic points if your claim concerns the whole day. A statement about a sustained pattern requires a suitable sequence, not a carefully selected pair. When a value is missing, leave the gap visible instead of filling it with zero. Zero would become an invented observation and could distort the apparent direction of a graph.
Historical downloads have a different purpose
NEA’s historical one-hour PM2.5 dataset is also available through data.gov.sg. At review, its catalogue covered April 2014 to September 2026, with five regional columns. That makes it a historical resource; it is not the place to obtain a current October reading. Check the coverage period before assuming a recently updated catalogue contains today’s data.
An update date can refer to the file or catalogue rather than the final observation in the file. Open the data and inspect its actual date range. A project comparing months should declare the start and end dates used, as well as any missing observations. This gives another reader a chance to reproduce the comparison rather than rely on a chart with an unexplained timeframe.
For a small personal spreadsheet, keep an untouched copy of the downloaded data and perform calculations in a separate sheet. Name columns with the metric and unit rather than “air quality.” If dates are imported as text, check the sorting before drawing a line chart. A sequence arranged alphabetically can look convincing while placing observations in the wrong order.
When comparing a saved download with a screenshot, also check the timezone. A computer may display a timestamp in its own local zone. Record Singapore time explicitly so that a chart viewed from another country does not appear to describe a different hour or calendar date.
A straightforward note for a group chat
Try this original message structure: “Source checked at [time]. Information relates to [region] and [observation or forecast period]. Metric: [label and unit]. Please use the linked official advisory for your circumstances. Next review: [time or before the activity].” Fill it only with information actually checked. The template is an organisational aid, not an official alert.
If you are correcting an earlier message, identify the correction directly. For example, say that the previous screenshot showed a forecast rather than an observation. Do not quietly replace an attachment and assume everyone will revisit it. A short explanation helps recipients understand whether the plan changed or only the description of the evidence changed.
Avoid adding a dramatic claim that the source does not make. A map of pollution readings does not, on its own, establish a precise source of smoke, a future duration or an impact on an individual. Those questions need their own evidence. Linking the official bulletin gives readers a way to distinguish an agency’s stated assessment from someone’s interpretation of a chart.
If you use the public API
The official developer documentation allows a date or date-time filter in Singapore time; without a date parameter, the PM2.5 endpoint returns its latest reading. The documented response includes regional metadata, and the unit is micrograms per cubic metre. These details help a small dashboard preserve the meaning of the source. They do not turn a home-built display into a new official advisory service.
Keep a visible source link and observation timestamp on any display you build. If a request fails, show that the current update is unavailable rather than replacing the previous value with zero or leaving an old value labelled “now.” If you retain an older value for context, label its age. A failure state should tell the user what is known and what has not refreshed.
For a chart, use explicit axis labels and avoid mixing PSI and concentration on one unlabeled scale. A technical demonstration can use clearly marked sample numbers, but those must not be presented as Singapore’s current air quality. This article has not implemented or validated a live dashboard; the suggestions are design and interpretation checks for anyone using the documented data.
Keep indoor-air questions separate
Regional outdoor information and household equipment answer different questions. A public chart cannot certify the air in your room, and a home device’s display should not be substituted for the official regional advisory. Our [household haze and indoor-air guide](https://monetizearticle.com/article/singapore-haze-and-indoor-air-a-practical-plan-for-your-home-ecc4c0) covers practical home planning separately. This article focuses on reading and communicating the data.
That separation helps avoid a misleading shopping conclusion. A change on a regional chart does not establish which purifier a household should buy, how it performed or what a particular person should do medically. A useful data explanation should make its scope clearer, not turn every number into a product recommendation.
Conclusion
Read the label, location and time before interpreting the number. Distinguish a one-hour concentration from a 24-hour index, an observation from a forecast and a fresh page load from a fresh measurement. Use the official advisory for the decision and person concerned. Those habits make a haze update more useful and reduce the chance that an old screenshot, incomplete chart or hypothetical example becomes mistaken for current official information.
Sources and research note
Checked 26 October 2026, Asia/Dhaka: [NEA air-quality FAQ](https://www.nea.gov.sg/our-services/pollution-control/air-pollution/faqs), [dated24October advisory](https://www.nea.gov.sg/media/news/advisories/index/haze-situation-update-24-october-2026), [MOH haze guidance](https://www.moh.gov.sg/others/haze/), [NEA PM2.5 catalogue](https://data.gov.sg/datasets/d_e1058d6974c877257e32048ab128ad83/view), [official API documentation](https://developers.data.gov.sg/datasets?formats=API&resultId=d_e1058d6974c877257e32048ab128ad83) and [historical one-hour data](https://data.gov.sg/datasets/d_6c0d5fc34838b12472475fdb73c0af29/view). All example messages and comparison numbers are original and hypothetical. Limited phrase searches do not establish comprehensive plagiarism clearance. The original AI-generated image is conceptual, not a current measurement or tested appliance. Saved Singapore Trends discovery is historical; this related guide is not a new independent trend claim.