Predict your own light bill
You just learned to forecast a stock. Here is a prediction that actually saves your family money: next month's JPS light bill. A bill is far calmer than a stock, so the exact same techniques, the moving average and the trend line, work beautifully. Grab six recent bills and let us forecast the next one.
JPS is the Jamaica Public Service Company, the electricity provider for most of Jamaica. Every month your household gets a bill showing how many units of energy you used (measured in kWh, kilowatt-hours) and the amount you owe in Jamaican dollars. Those two numbers are all we need.
Your step-by-step guide
- Gather six recent bills. Ask a parent or guardian for the last six months of JPS bills (paper or the online account). From each one, note two numbers: the units used (kWh) and the total amount (J$).
- Line them up oldest to newest. Order matters for a trend. Bill 1 is the oldest, Bill 6 is the most recent.
- Method 1, the moving average. Add the last three amounts and divide by three. That is a calm, steady guess for next month, it assumes things stay about the same.
- Method 2, the trend line. Fit a straight line through all your amounts and extend it one month. If your bills have been climbing, this catches the rise the average misses.
- Method 3, AR (auto-regression). Learn how each bill relates to the one just before it, then build the next bill from your most recent one. It leans on your latest reading.
- Blend and sanity-check. The truth usually sits among the three. Look at the chart, read the comment, and pick your final number.
The calculator
The rows start blank. Enter at least three recent bills, oldest first, and press ＋ Add another bill for more, then press Predict next month. No bills handy? Press Use example to see how it works. Everything runs in your browser, nothing is sent anywhere.
| Bill | Units used (kWh) | Amount (J$) |
|---|
A single forecast number is a guess, and no guess is ever exactly right. So next to the dashed forecast line, the chart draws a shaded error band, also called a confidence interval. Think of it as the calculator saying, "I expect the bill somewhere in here," instead of pretending to know the exact cent.
The band is built from how much your past bills wobbled around their trend. Calm, steady bills give a narrow band, so you can trust the number. Jumpy bills give a wide band, a signal to plan for a range, not a single figure. Statisticians usually size the band so the real value falls inside it about 19 times out of 20. A weather app does the same thing when it says "a high near 32°C, give or take two degrees."
Jamaica is in a heatwave right now, and heat is the number one reason light bills jump: fans and air conditioning run for hours, and fridges work harder. If your forecast looks scary, that is useful information. A few habits genuinely lower the next bill: set the AC no lower than 24°C, use fans first, switch off lights and devices you are not using, and run the washing machine with full loads. Predicting the bill is step one; changing it is step two.
- The same tools that forecast a stock, moving average, a trend line, and AR, forecast a bill.
- On a calm, rising series the trend line reads higher; the average is more cautious.
- The error band shows the likely range around your forecast: narrow when bills are steady, wide when they jump.
- Predicting your bill is the first step; saving electricity is how you change it.
You have put every technique from the course to work on a real bill. Well done.