A derivative is a rate
A derivative tells you how fast one quantity changes relative to another at a particular moment. Speed is the best-known case: it is how fast position changes with time. If you can read a speedometer, you already understand the idea. The calculus adds a way to compute it from a formula or a graph.
Average rates first
An average rate of change is the change in a quantity divided by the change in the other quantity.
If a vehicle covers 100 m in 20 s, its average speed is 100 / 20 = 5 m/s. On a graph of position against time, that is the slope of the straight line joining the start and end points: rise (100 m) over run (20 s).
Slope is the unifying picture. Whenever you see “per” in a unit (metres per second, volts per metre, litres per minute), there is a slope behind it.
From average to instantaneous
Average speed hides what happens inside the interval. A vehicle can accelerate, cruise and brake within those 20 s. To find how fast it moves at one instant, take a shorter and shorter interval around that instant.
Take a simple illustration where position (in metres) is x = t², with t in seconds. At t = 1 s the position is 1 m. Now compute average speeds starting from t = 1:
| Interval | Position change | Time change | Average speed |
|---|---|---|---|
| 1 s to 3 s | 9 − 1 = 8 | 2 | 4 m/s |
| 1 s to 2 s | 4 − 1 = 3 | 1 | 3 m/s |
| 1 s to 1.1 s | 1.21 − 1 = 0.21 | 0.1 | 2.1 m/s |
As the interval shrinks, the average settles toward 2 m/s. That limit is the derivative at t = 1: the instantaneous speed, or the slope of the tangent line, which is the straight line that just touches the curve at that point.
A rule that gives the answer directly
For powers there is a pattern. The derivative of tⁿ is n × tⁿ⁻¹. For x = t², the derivative is 2t. At t = 1 that gives 2 m/s, matching the shrinking-interval table. At t = 3 it gives 6 m/s.
Some other rules that cover many cases:
- The derivative of a constant is 0, since a constant does not change.
- The derivative of a straight line y = m x + b is m, its slope.
- Constants multiply through: the derivative of 5t² is 5 × 2t = 10t.
- Derivatives of sums are sums of derivatives.
Software and symbolic tools handle more complicated functions, and you do not need to memorise much to use the concept well.
Chains of rates
Rates of change can be stacked. Position changes with time at a rate called velocity, and velocity changes with time at a rate called acceleration.
Continuing the example, x = t² gives velocity v = 2t. The derivative of 2t is 2, so the acceleration is a steady 2 m/s². Check: velocity is 2 m/s at t = 1 and 4 m/s at t = 2, a gain of 2 m/s per second. It agrees.
Notice the units. Position in m, velocity in m/s, acceleration in m/s². Each derivative divides by one more unit of time. Tracking this is a simple application of dimensional analysis and helps catch mistakes.
Reading slope from a graph
You do not always have a formula. With a graph, estimate the derivative by drawing a short straight line along the curve at the point of interest and computing rise over run, using the axes’ units.
Suppose a plot of tank volume against time rises from 40 L to 70 L over 10 minutes along a nearly straight stretch. The rate is (70 − 40) / 10 = 3 L/min. If the curve is steeper elsewhere, the rate there is larger. A flat stretch means a rate of zero: nothing is changing. A downward slope means a negative rate.
This is how to interpret many datasheet graphs, such as how quickly a quantity falls as temperature rises. The slope tells you the sensitivity.
Where derivatives show up
- Speed and acceleration from position data.
- Current in a capacitor, which depends on how fast voltage changes. A steadily changing voltage gives a steady current; a constant voltage gives none.
- Heating or cooling rate of an object, in degrees per minute.
- Gradient of a calibration curve, which gives sensitivity.
- Optimisation: at the top or bottom of a smooth curve the slope is zero, which helps locate maxima and minima.
The reverse operation
Derivatives and integrals undo each other. If you know how fast something changes at every moment, the total accumulated change comes from the area under that rate graph. Speed to distance is the standard example. The areas under curves refresher develops that side.
Practical cautions
Real measurements are noisy, and a derivative magnifies noise because it compares nearby values. Estimating a rate from two readings taken very close together can give wildly erratic results. Smoothing the data or using a wider interval often helps, at the cost of detail. Remember also that every reading carries measurement uncertainty, which affects a computed slope.
Always include units when quoting a rate, and compare against a rough expectation as one of your sanity checks. These examples are simple illustrations with chosen numbers. Where a calculation supports a safety or regulatory decision, use the methods required in your field and have it reviewed by a qualified person.