Estimating Derivatives of a Function at a Point
Why this matters
Imagine you're on a road trip from Dallas to Atlanta. Your friend is driving, and you're tracking the car's progress. You don't have a working speedometer, but every 15 minutes, you jot down the total miles traveled. At the three-hour mark, you wonder, "Exactly how fast were we going at this very moment?" You can't know for sure. But you can make a really good estimate.
You could calculate your average speed between 2 hours 45 minutes and 3 hours. Or your average speed between 3 hours and 3 hours 15 minutes. Better yet, you could average those two speeds together to get a sharp estimate for your speed at the 3-hour instant.
This is exactly what we're doing in calculus when we estimate derivatives. We're finding the instantaneous rate of change when we don't have a perfect function, just clues from a table or a graph. Let's learn how to turn those clues into solid estimates.
Concept overview
flowchart TD
A[Start: Estimate f'(c)] --> B{What information do you have?};
B --> C{Function f(x)};
B --> D{Table of values};
B --> E{Graph of f(x)};
C --> C1[Use a calculator's nDeriv function];
D --> D1[Find smallest interval around c, like (a, b)];
D1 --> D2[Calculate slope: (f(b) - f(a)) / (b - a)];
E --> E1[Draw the tangent line at x=c];
E1 --> E2[Pick two points on the tangent line];
E2 --> E3[Calculate slope between those two points];
C1 --> F[Result is the estimated derivative];
D2 --> F;
E3 --> F;
Core explanation
Welcome back! In the last few topics, we've been building up the formal definition of the derivative. Now, we're going to get practical. The AP exam and the real world won't always hand you a nice, clean function like f(x) = x^2. Sometimes, you'll only have data points in a table or a picture of a graph.
Your goal is still the same: find the derivative, which means finding the slope at one specific point. But now, we have to estimate it. Let's break down the three ways you'll be asked to do this.
1. Estimating from a Table of Values
This is the most common scenario you'll see. Imagine a table that shows the temperature of a pizza, P(t), in degrees Fahrenheit, t minutes after it comes out of the oven.
t (minutes) |
0 | 2 | 5 | 9 | 12 |
|---|---|---|---|---|---|
P(t) (°F) |
450 | 380 | 290 | 210 | 180 |
Let's say we need to estimate the rate of change of the temperature at t = 5 minutes. In calculus terms, we want to estimate P'(5).
We can't use the limit definition of the derivative because we don't have a function. But we know the derivative is just a slope. So, we can calculate the slope of a line that's really close to the tangent line at t = 5. This is the slope of the secant line.
You have a few options:
- Use the interval just before the pointCalculate the slope between
t=2andt=5.Slope = (290 - 380) / (5 - 2) = -90 / 3 = -30 °F/min - Use the interval just after the pointCalculate the slope between
t=5andt=9.Slope = (210 - 290) / (9 - 5) = -80 / 4 = -20 °F/min
Both are reasonable estimates. But which one is best?
The best estimate comes from the smallest possible interval that contains your point. In this case, the interval from t=2 to t=9 contains t=5.
So, for t=5, the best estimate uses the points at t=2 and t=9.
P'(5) ≈ (P(9) - P(2)) / (9 - 2) = (210 - 380) / 7 = -170 / 7 ≈ -24.286 °F/min
This method is called the symmetric difference quotient when the points are equally spaced. It's like our road trip example: we're averaging our speed from a little before our target time and a little after. It gives us a more balanced picture.
2. Estimating from a Graph
What if you're given a graph of a function, say y = g(x), and asked to estimate g'(1)?
Here, the derivative g'(1) is the slope of the tangent line to the curve at x = 1. Your job is to be an artist and a mathematician.
- 1Find the pointLocate
x = 1on the graph. - 2Draw the tangent lineTake a straightedge (like a ruler or the side of your ID card) and carefully draw a line that just "kisses" the curve at that one point. The line should match the steepness of the curve at that exact spot.
- 3Calculate the slope of your lineNow, ignore the curve and focus on the straight line you just drew. Pick two points on your line that are easy to read from the grid. Let's say your line appears to go through
(0, 1.5)and(3, 0). - 4Use the slope formula
m = (y₂ - y₁) / (x₂ - x₁)g'(1) ≈ (0 - 1.5) / (3 - 0) = -1.5 / 3 = -0.5
Because you are drawing the line, your estimate might be slightly different from a classmate's. That's okay! On the AP exam, they provide a range of acceptable answers for these questions. The key is to show your work by listing the two points you used to calculate the slope of your tangent line.
3. Using Technology
Finally, the easiest method! Your graphing calculator is a powerful tool for estimating derivatives. If you are given a function, like f(x) = x * cos(x), and asked for f'(2), you can use your calculator's built-in numerical derivative tool.
On a TI-84 calculator, the steps are:
- Press the
MATHbutton. - Scroll down to
8: nDeriv(and pressENTER. - You'll see a template that looks like this:
d/dx ( ) | x = - Fill it in:
- The variable is
x. - The function is
x * cos(x). - The point is
x = 2.
- The variable is
Your screen will look like this:
d/dx (x*cos(x)) | x = 2
Press ENTER, and the calculator will give you the estimate: ≈ -2.728.
The calculator isn't doing magic. It's using a very, very precise version of the symmetric difference quotient with a tiny value for h (like h = 0.001). It calculates (f(2.001) - f(1.999)) / 0.002 to get an incredibly accurate estimate.
So, whether you have a table, a graph, or a function with a calculator, you have a strategy to find the instantaneous rate of change.
Worked examples
Let's walk through a couple of examples together so you can see these methods in action.
Estimating from a Table
The population of a suburban town, P, is modeled by a differentiable function of time t, where t is years since 2010. The table below gives values of P(t) for selected years.
t (years) |
0 | 3 | 4 | 7 | 9 |
|---|---|---|---|---|---|
P(t) (people) |
12,500 | 13,100 | 13,400 | 14,200 | 14,900 |
Problem: Estimate the rate at which the population was changing in the year 2014. Show the computations that lead to your answer and indicate units of measure.
Solution Walkthrough:
- 1Identify the GoalThe question asks for the "rate at which the population was changing" in 2014. This is code for the derivative. The year 2014 corresponds to
t = 4. So, we need to estimateP'(4). - 2Locate the Point in the TableFind
t = 4in our table. The data points immediately surrounding it are att = 3andt = 7. - 3Choose the Best MethodSince we have data on both sides of
t = 4, we should use the average rate of change over the interval[3, 7]. This will give us the most balanced estimate. - 4Perform the CalculationWe'll use the slope formula with the points
(3, 13100)and(7, 14200).P'(4) ≈ (P(7) - P(3)) / (7 - 3)P'(4) ≈ (14200 - 13100) / 4P'(4) ≈ 1100 / 4P'(4) ≈ 275 - 5State the Final Answer with UnitsThe units for a derivative are always
(units of y) / (units of x). Here, that'speople / year. Answer: The population was changing at an estimated rate of 275 people per year in 2014.
Estimating from a Graph
The graph of a function f(x) is shown below.
(Imagine a smooth, curvy graph. Let's say it passes through (-2, 3), has a local max near x=-1, passes through (0, 1), has a local min near x=1.5, and passes through (3, 0).)
Problem: Estimate the value of f'(0).
Solution Walkthrough:
- 1Identify the GoalWe need to find the slope of the tangent line at
x = 0. - 2Locate the Point and Draw the TangentFind
x = 0on the graph. The point is(0, 1). Now, take a straightedge and draw a line that just touches the curve at(0, 1)and has the same slope as the curve at that point. The curve is decreasing atx=0, so our line should have a negative slope. - 3Find Two Points on Your Tangent LineLook carefully at the line you just drew. Find two points on it that are easy to read. Let's say your line appears to pass perfectly through
(0, 1)and(2, -1). (Note: Your points might be slightly different, like(-1, 2)and(1, 0). That's okay! The process is what matters.) - 4Calculate the SlopeUse the slope formula with the two points you chose from your tangent line.
f'(0) ≈ (y₂ - y₁) / (x₂ - x₁)f'(0) ≈ (-1 - 1) / (2 - 0)f'(0) ≈ -2 / 2f'(0) ≈ -1 - 5State the Final AnswerThe estimated value of
f'(0)is -1.
Try it yourself
Ready to try a couple on your own? Don't worry about getting the perfect answer; focus on applying the right process.
Problem 1 (Table):
Maya is flying a drone. The drone's altitude A(t) in feet at time t in seconds is given in the table below.
t (seconds) |
0 | 10 | 12 | 15 | 20 |
|---|---|---|---|---|---|
A(t) (feet) |
5 | 60 | 72 | 90 | 110 |
Estimate A'(12). What are the units of your answer?
Hint: What's the smallest interval in the table that contains t=12? Use the endpoints of that interval to calculate the slope.
Problem 2 (Graph):
Consider the graph of h(x) below. Estimate h'(-2).
(Imagine a graph that goes through (-4, 0), has a sharp corner (a "cusp") at (-3, 2), goes through (-2, 1), and has a horizontal tangent at (0, -1).)
Hint: Carefully draw a tangent line at x = -2. Then, find two points on that line and calculate the rise over run. What do you notice about the slope at x = -3? Can you estimate it?
In simple terms, this topic is about how to make a smart guess for the instantaneous rate of change (the derivative) at a single point when you only have a table, a graph, or a calculator.
d/dx (x*cos(x)) | x = 2
- CHA-2.D: Estimate derivatives.
- CHA-2.D.1
- The derivative at a point can be estimated from information given in tables or graphs.
- CHA-2.D.2
- Technology can be used to calculate or estimate the value of a derivative of a function at a point.
flowchart TD
A[Start: Estimate f'(c)] --> B{What information do you have?};
B --> C{Function f(x)};
B --> D{Table of values};
B --> E{Graph of f(x)};
C --> C1[Use a calculator's nDeriv function];
D --> D1[Find smallest interval around c, like (a, b)];
D1 --> D2[Calculate slope: (f(b) - f(a)) / (b - a)];
E --> E1[Draw the tangent line at x=c];
E1 --> E2[Pick two points on the tangent line];
E2 --> E3[Calculate slope between those two points];
C1 --> F[Result is the estimated derivative];
D2 --> F;
E3 --> F;
Read what Saavi narrates
Hi everyone, it's Saavi. Let's talk about one of the most practical skills in calculus: estimating derivatives.
Think about this: you're on a long drive, and you want to know your exact speed at the 3-hour mark. But your speedometer is broken. All you have is a log of your total mileage every 15 minutes. You can't know your speed for sure... but you can make a really good guess, right? You could calculate your average speed for the 15 minutes right before 3 hours, and for the 15 minutes right after. That's exactly what we're doing today.
The derivative is the instantaneous rate of change, or the slope at one single point. When we don't have a perfect equation, we have to estimate it using clues from a table or a graph.
Let's try one with a table. Imagine a town's population, P, is measured every few years. Let's say at year 3, the population is 13,100... and at year 7, it's 14,200. We want to estimate the rate of population growth in year 4.
To get the best estimate for the rate at year 4, we should use the data we have on either side of it. So we'll use the points for year 3 and year 7. We just calculate the slope between them.
The change in population is 14,200 minus 13,100... which is 1,100 people.
The change in time is 7 minus 3... which is 4 years.
So the rate is 1,100 divided by 4, which gives us 275 people per year. That's our estimate for the derivative at year 4.
Now, here's a really common mistake. A lot of students see a question asking for the rate of change at a point, say at time equals 4, and they just look at the table and report the population value at time equals 4. Please don't do this! The derivative is a *rate*, a *slope*. It's not the y-value. You always need to calculate a slope between two points.
You've got this. The key is to remember that the derivative means slope. Whether you're looking at a table or a graph, your job is to find the slope of the line that best represents the rate of change at that single point. Keep practicing, and it'll become second nature.
The derivative `f'(c)` is the *rate of change* (slope) at `c`, not the *value* (y-coordinate) at `c`.
Always calculate a slope using the formula `(y₂ - y₁) / (x₂ - x₁)`.
This only gives you the rate of change *after* the point. A more accurate estimate comes from averaging the behavior before and after.
Use the closest points on *both sides* of the point of interest. In the example, use the points at `x=4` and `x=6` to calculate the slope: `(f(6) - f(4)) / (6 - 4)`.
This calculates the slope of a *secant line*, which is just an approximation. The goal is to find the slope of the *tangent line*.
First, draw the tangent line. Then, pick two convenient points that lie *on that straight line* to calculate the slope.
Rates of change have context. Forgetting units can lose you a point.
Always write the units as (units of y) / (units of x). For example, `gallons per minute`, `dollars per year`, or `feet per second`.
This interval is far too wide and will produce a poor estimate.
Always use the *smallest possible interval* that contains your point. In this case, that's `[5, 10]`.