Preparing a workbook
Use an XLSX workbook with the observations you want to compare on the same sheet. Choose the header that identifies the columns, then select the time or other X column and the Y columns containing the measurements. For example, one column might contain observation times and two others might contain temperatures from different sensors. Selecting both measurement columns lets you compare them against the same X values without mixing the original measurements with their rates of change.
Before interpreting the charts, inspect the source sheet. Make sure the selected header belongs to the data, the X values represent the intended time or quantity, and each Y column has the expected measurements. Duplicate timestamps or X values and invalid points are omitted; missing or unusable observations can leave a less complete picture than the original table. Check the workbook when a chart contains fewer points than expected.
Reading the charts
The chart shows the selected values against X. Solid lines are the original measurements on the left axis, useful for seeing whether a sensor reading rises or falls over the recorded period. Dashed lines are estimates of the first derivative on the right axis: how quickly each value changes as X changes. A positive rate indicates an increase, while a negative rate indicates a decrease. A rate near zero suggests little change over the observations used for that estimate. The two are combined in one dual-axis chart by default; clear the merge option to show them as two separate charts.
The vertical axes use different units, so a rate is not directly comparable to an original measurement. Do not compare the heights of a solid and a dashed line, and read exact values from the hover readout or the table below the chart.
The rates are calculated using numerical differences between data points, not by fitting a quadratic curve. They approximate change in the observed data; they are not exact derivatives of an underlying process. Nonuniform timestamps can be used because the actual X differences matter. A change of ten units over two hours is a different rate from the same change over ten hours. Duplicate X values do not provide a usable interval and are omitted, as are invalid points.
The rates also offer smoothing. Selecting 3, 5, 9 or 15 points shows, at each position, the centered average of the rates at that many neighboring samples. A wider window gives a smoother curve but can weaken brief real changes. Smoothing applies only to the rate curves; the original values are always shown unsmoothed, and smoothing does not restore missing or omitted points.
Check units and uncertainty
A derivative’s unit is the Y unit divided by the X unit. For example, a temperature measured in degrees against time measured in hours gives a rate in degrees per hour. Verify what the workbook’s timestamps and measurements mean before comparing rates or reporting them. If series use different measurement units, their values and rates are not directly comparable without accounting for those units.
Small changes in measurements can become conspicuous when numerical differences are taken, particularly over short intervals. Noise, rounding, gaps and omitted points can affect the estimated rate, so investigate unexpected spikes in the source data rather than assuming they describe a real event. The charts support exploration; important conclusions should be checked against the original observations and the context in which they were collected.