How to Analyse Punggol Haze Data in a Spreadsheet: Start Here

How do you create a trustworthy Singapore haze chart in Excel or Google Sheets? Start with data that really belong to the North reporting region containing Punggol, preserve the timestamp and physical unit, calculate simple statistics and a clearly labelled rolling mean, and explain the limits of what the analysis can prove. Do not turn a spreadsheet average into Singapore’s official 24-hour PSI. The PSI uses a separate defined methodology involving pollutants and sub-indices.

The first part of this guide uses twelve invented one-hour PM2.5 concentrations so that learners can practise calculations safely without confusing practice with today’s outdoor conditions. A later section explains how to find NEA data on data.gov.sg, import a historical CSV, and preserve the source metadata. For immediate health decisions, always use NEA’s current haze service, not a static file or student chart.

Readers searching “Singapore haze Excel tutorial”, “Google Sheets PM2.5 moving average”, “how to graph air-quality data” or “Punggol PSI spreadsheet” need a method they can actually inspect. This article provides formulas, manual checks, missing-value decisions, interpretations, common mistakes and a self-testing workflow. A good chart should be understandable, reproducible and honest about uncertainty—not merely dramatic.

Why the North Region Matters

NEA reports the one-hour PM2.5 concentration and 24-hour PSI across North, South, East, West and Central. Its air-pollution FAQ explicitly includes Punggol under North. A regional reading is not a certified monitor attached to a particular park bench, school gate or apartment.

A Punggol resident may choose to analyse North and compare it with another region, but must not quietly rename the North data “Punggol Waterway sensor”. The unit, time and reporting geography determine what comparison is defensible. A well-labelled line chart can help people understand regional change without falsely advertising block-level precision.

There is also a difference between looking at a real official historical series and a purely fictional worksheet. Fictional values can make the arithmetic teachable and safe; official values require a source, date, time-zone interpretation and data-handling notes. Both uses can be legitimate when clearly labelled.

Choose a Source That Supports the Question

The NEA PM2.5 dataset directory lists real-time and historical resources. The historical one-hour PM2.5 CSV offers a convenient entry point for classroom analysis. The data.gov.sg real-time API guide describes the API endpoint and query options for live retrieval.

A historical CSV is useful when a student needs stable data that can be reproduced next week. A live API is useful for authorised software development, but data availability, time zones and rate limits still need management. Neither one automatically supplies a room’s air concentration or a personal medical conclusion.

For Python and JSON extraction, the site already has a separate Punggol Haze API tutorial. This page focuses on cell formulas, chart design, data validation, mathematical interpretation and honest reporting.

Make a Data Dictionary Before Typing

A simple dataset benefits from explicit column names: Hour index, PM2.5 (µg/m³), Three-hour trailing mean, One-hour change, Percentage change and Data quality note. For real records, add the source timestamp, region and retrieval date.

A data dictionary says how each column is defined. For instance, “one-hour change” can mean the latest row minus the previous consecutive hourly observation. If a timestamp is missing, the time gap may be longer than one hour, and the formula may need to be adjusted. “Three-hour trailing mean” should require three valid observations rather than silently average whatever happens to remain.

Think of the data dictionary as part of the Mathematics. An unlabeled value of 60 could mean an index, minutes, a physical concentration, an appliance wattage or a classroom head count. The numerical operation may be right but meaningless if the quantity is not named.

The Original Fictional Twelve-Hour Practice Series

Enter hour index 1–12 in cells A2:A13 and fictional PM2.5 concentrations 25, 35, 42, 48, 55, 70, 65, 50, 45, 40, 32, 28 in cells B2:B13. Use B1 for the header PM2.5 (µg/m³) and A1 for Hour index.

The sequence rises to a peak at index 6 and then declines; that shape makes it suitable for comparing raw values with rolling means. These are not NEA readings on any real 2026 date. Put the phrase “Fictional data for learning, not live Punggol air quality” visibly above the table and beneath every exported chart.

For a second exercise, replace the hour index with actual consecutive timestamps from a properly sourced archive. Preserve the original data in a “Raw” sheet and calculate derived values in a separate “Analysis” sheet so that another student can audit what was changed.

Worked Task 1: Count All Numeric Readings

Question. How many numerical readings are in B2:B13?

Solution. Use =COUNT(B2:B13); the result should be 12.

Reasoning and correction. COUNT ignores text. If the visible table has twelve rows but COUNT reports eleven, investigate an entry imported as text or a missing measurement. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 2: Find the Sum

Question. What is the sum of all twelve fictional values?

Solution. =SUM(B2:B13) returns 535.

Reasoning and correction. The sum of concentration values is not automatically a cumulative personal dose and certainly not a PSI index. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 3: Calculate the Mean

Question. Calculate the arithmetic mean of the twelve hourly concentrations.

Solution. =AVERAGE(B2:B13) returns 44.5833… µg/m³, or about 44.58 µg/m³.

Reasoning and correction. This is a twelve-observation mean. It is not a 24-hour PSI and not a current NEA measurement. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 4: Find the Median

Question. What is the middle concentration after ordering all twelve numbers?

Solution. Sorted numbers are 25,28,32,35,40,42,45,48,50,55,65,70. =MEDIAN(B2:B13) returns 43.5 µg/m³.

Reasoning and correction. With an even number of entries, take the mean of the sixth and seventh sorted values: (42+45)/2. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 5: Find the Maximum

Question. Locate the highest fictional hourly concentration.

Solution. =MAX(B2:B13) returns 70 µg/m³, at index 6.

Reasoning and correction. A chart maximum applies only to the selected time period, not the whole year or a specific Punggol estate. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 6: Find the Minimum

Question. Locate the lowest value.

Solution. =MIN(B2:B13) returns 25 µg/m³, at index 1.

Reasoning and correction. The lowest number in one short sample is not a statement that the hour was medically safe. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 7: Calculate the Range

Question. Find the difference between maximum and minimum.

Solution. =MAX(B2:B13)-MIN(B2:B13) returns 45 µg/m³.

Reasoning and correction. Range describes spread between extremes but ignores how long each value persisted. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 8: Find an Hourly Change

Question. Find the change from the first value 25 to the second 35.

Solution. In D3 (the change column), use =B3-B2; the result is +10 µg/m³.

Reasoning and correction. The positive sign means the later fictional concentration is higher. It does not identify a source fire. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 9: Calculate Percentage Increase

Question. Express the change from 25 to 35 as a percentage of the earlier value.

Solution. =(B3-B2)/B2 displays 40% when formatted as a percentage.

Reasoning and correction. The denominator must be the earlier value. Percentage concentration change does not equal percentage health-risk change. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 10: Calculate Percentage Decrease

Question. The sample drops from 70 at index 6 to 65 at index 7. What percentage change is this?

Solution. (65−70)/70×100 ≈ −7.14%.

Reasoning and correction. A drop after a peak cannot prove that every regional fire stopped. It is a comparison of values. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 11: Calculate a Trailing Three-Hour Mean

Question. What is the first complete mean of 25, 35 and 42?

Solution. (25+35+42)/3 = 34 µg/m³. In C4, use =IF(COUNT(B2:B4)=3,AVERAGE(B2:B4),"").

Reasoning and correction. Require all three observations. A rolling mean is not an official PSI computation. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Worked Task 12: Move the Window Forward

Question. Find the next trailing mean for 35, 42 and 48.

Solution. 125/3 ≈ 41.67 µg/m³. Copy the C4 formula down to C5.

Reasoning and correction. Two adjacent rolling windows share two observations, so their values are statistically related by construction. Before viewing the answer, the learner should identify the range and variable being measured. After calculating, they should change one input or shift the window by one row and explain exactly which numbers now contribute. A copied formula is not mastered until the student can reproduce the correct result independently.

Evidence boundary. The entire numerical series is fictional. A precisely formatted cell is not a health advisory or a certified measurement. To use real historical data, replace the values with a dated NEA source, retain its timestamp and reporting region, and document the method.

Build Your First Correct Chart

Select the hour index and concentration columns, then insert a line chart. Use Hour index on the horizontal axis and PM2.5 concentration (µg/m³) vertically. The chart title should say Fictional Twelve-Hour PM2.5 Exercise so readers cannot confuse it with a live Punggol report.

When adding a trailing three-point mean, treat it as a separate data series. The mean starts only when three complete observations exist, so do not insert invented zeroes into the first two positions. Include a legend and a caption explaining that the smoothed series may hide short spikes.

A good chart can be understood without someone explaining it orally: it has title, labels, units, period and provenance. The objective is not to make every change look dramatic; it is to show what the fictional values actually do.

Why Your Y-Axis Scale Changes the Story

The same data can look visually calm or alarming depending on whether the y-axis begins at zero or a value close to the minimum. A narrow range magnifies fluctuations; a broad range makes the variation appear smaller.

For a column chart, truncating the baseline can seriously distort perceptions of relative size. For a line chart focused on change, a narrower scale may be justified but should be clearly labelled. Students can create two versions and write an explanation of how presentation choices influence readers.

The underlying numbers did not change. That is the key media-literacy lesson. A responsible author does not select an axis purely to manufacture a headline about Punggol haze. Both scale and purpose need to be disclosed.

Why an Hourly Plot Is Not an Exposure Prescription

An hour-index plot does not describe the actual concentration at every indoor space, the person’s minute-by-minute breathing, or the relevant health group. Regional measurements and individual exposure are different concepts.

Even a statistically careful plot can only answer the specific question supported by its data. For instance, the maximum in the practice series is 70 µg/m³. That does not tell whether a child in Punggol may take part in a long outdoor sports session today.

Refer readers to NEA’s personal guide for public decisions and to individual medical guidance when relevant. A spreadsheet is a learning and analysis tool, not a substitute for the responsible public agency.