Partial Derivative Calculator: Step-by-Step Guide with Worked Examples
A partial derivative calculator finds the rate of change of a multivariable function with respect to one variable at a time, holding every other variable constant. Partial derivatives show up the moment a function depends on more than one input — temperature that varies with both position and time, profit that depends on price and quantity, or a surface z = f(x, y) where you need the slope in the x-direction separately from the slope in the y-direction. This guide walks through the notation, the core differentiation rules applied to multivariable functions, second-order and mixed partial derivatives, the mistakes that trip up students most often, and a full set of practice problems with checked answers.
Contents
- 01What Is a Partial Derivative Calculator and When Do You Need One?
- 02How Does a Partial Derivative Calculator Work Step by Step?
- 03Core Rules for Differentiating Multivariable Functions
- 04How Do You Find Second-Order Partial Derivatives?
- 05Common Mistakes When Computing Partial Derivatives
- 06Practice Problems: Partial Derivatives with Full Solutions
- 07Why Do Partial Derivatives Matter in Real Applications?
- 08What Are Common Questions About Partial Derivative Calculators?
What Is a Partial Derivative Calculator and When Do You Need One?
A partial derivative calculator computes ∂f/∂x, ∂f/∂y, or ∂f/∂z for a function of several variables, treating every variable except the one you are differentiating with respect to as a fixed constant. The symbol ∂ (a rounded 'd') distinguishes a partial derivative from an ordinary derivative d/dx, which only makes sense for single-variable functions. Consider z = f(x, y) = x²y + 3y². This surface has a different slope depending on which direction you move. ∂f/∂x tells you the slope if you move in the x-direction while y stays fixed. ∂f/∂y tells you the slope if you move in the y-direction while x stays fixed. Both are needed to fully describe how the surface behaves near any point — neither one alone is enough. You need partial derivatives whenever a quantity depends on two or more independent variables: heat distribution across a metal plate (depends on x, y, and time t), a company's profit (depends on price and advertising spend), or a machine learning loss function (depends on every model parameter simultaneously). A partial derivative calculator step by step shows exactly which variables were held constant at each stage, which is the detail students most often skip when working by hand.
∂f/∂x means: differentiate f with respect to x, and treat every other variable as if it were a plain number. Nothing else about differentiation changes.
How Does a Partial Derivative Calculator Work Step by Step?
Whether you compute by hand or use a partial derivative calculator step by step, the process below applies to every multivariable function. Work through this worked example alongside each step: f(x, y) = x³y² + 4xy − y³. Find ∂f/∂x and ∂f/∂y, then evaluate both at the point (2, 1).
1. Step 1 — Decide which variable you are differentiating with respect to
Pick one variable to treat as the active variable. Every other letter in the expression becomes a constant for this pass, exactly like a number such as 5 or −3. For ∂f/∂x, y is frozen; for ∂f/∂y, x is frozen.
2. Step 2 — Apply ordinary differentiation rules to the active variable
Differentiate term by term using the Power Rule, Product Rule, or Chain Rule as usual, but keep the frozen variable exactly as written. For ∂f/∂x of x³y² + 4xy − y³: the term x³y² becomes 3x²y² (y² is just a constant multiplier), 4xy becomes 4y (x differentiates to 1, so 4·1·y = 4y), and −y³ becomes 0 (no x present, so it is a constant with respect to x). Result: ∂f/∂x = 3x²y² + 4y.
3. Step 3 — Repeat for each remaining variable
Now find ∂f/∂y, treating x as constant. The term x³y² becomes x³·2y = 2x³y, the term 4xy becomes 4x (y differentiates to 1, so 4x·1 = 4x), and −y³ becomes −3y². Result: ∂f/∂y = 2x³y + 4x − 3y².
4. Step 4 — Evaluate at the given point
Substitute x = 2, y = 1 into each partial derivative. ∂f/∂x(2,1) = 3(2)²(1)² + 4(1) = 3·4·1 + 4 = 12 + 4 = 16. ∂f/∂y(2,1) = 2(2)³(1) + 4(2) − 3(1)² = 2·8·1 + 8 − 3 = 16 + 8 − 3 = 21.
5. Step 5 — Check numerically
f(2,1) = (2)³(1)² + 4(2)(1) − (1)³ = 8 + 8 − 1 = 15. Nudge x only: f(2.001, 1) = (2.001)³ + 4(2.001) − 1 ≈ 8.012006 + 8.004 − 1 = 15.016006. [f(2.001,1) − f(2,1)] / 0.001 = 0.016006 / 0.001 ≈ 16.0 — matches ∂f/∂x = 16 ✓. Nudge y only: f(2, 1.001) = 8(1.002001) + 4(2)(1.001) − (1.001)³ ≈ 8.016008 + 8.008 − 1.003003 = 15.021005. [f(2,1.001) − f(2,1)] / 0.001 = 0.021005 / 0.001 ≈ 21.0 — matches ∂f/∂y = 21 ✓.
Core Rules for Differentiating Multivariable Functions
Every rule you already know from single-variable calculus — Power Rule, Product Rule, Chain Rule — still applies to partial derivatives. The only new habit is remembering which letters are frozen during each pass. Power Rule example: h(x, y) = x⁴y³. ∂h/∂x = 4x³y³ (y³ is a constant coefficient). ∂h/∂y = 3x⁴y² (x⁴ is a constant coefficient). Product Rule example: g(x, y) = x²·eʸ. There is no product between x-terms here, so ∂g/∂x = 2x·eʸ directly. But for k(x, y) = x²·ln(x) + y², finding ∂k/∂x requires the Product Rule on x²·ln(x): treat f = x², f' = 2x, and g = ln(x), g' = 1/x, giving 2x·ln(x) + x²·(1/x) = 2x·ln(x) + x. Then ∂k/∂x = 2x·ln(x) + x (the y² term vanishes since it has no x). For ∂k/∂y, x²·ln(x) is entirely constant with respect to y, so ∂k/∂y = 2y. Chain Rule example: p(x, y) = (3x + 2y)⁵. Treating y as constant, the outer function is u⁵ with inner u = 3x + 2y. ∂p/∂x = 5(3x + 2y)⁴ · 3 = 15(3x + 2y)⁴. Treating x as constant instead, the inner derivative with respect to y is 2, so ∂p/∂y = 5(3x + 2y)⁴ · 2 = 10(3x + 2y)⁴. Logarithm example: m(x, y) = ln(xy). Since ln(xy) = ln(x) + ln(y) by the log product rule, ∂m/∂x = 1/x and ∂m/∂y = 1/y — much faster than applying the Chain Rule directly to ln(xy).
The rules don't change — Power Rule, Product Rule, and Chain Rule all still apply. Only the bookkeeping changes: freeze every variable except the one you're differentiating.
How Do You Find Second-Order Partial Derivatives?
A second-order partial derivative differentiates an already-differentiated function one more time. Because there are two first partial derivatives (∂f/∂x and ∂f/∂y), there are four second-order partials: ∂²f/∂x² (differentiate ∂f/∂x with respect to x again), ∂²f/∂y² (differentiate ∂f/∂y with respect to y again), and two mixed partials ∂²f/∂x∂y and ∂²f/∂y∂x, which differentiate in opposite orders.
1. Worked example — f(x, y) = x³y² + 2xy
First partials: ∂f/∂x = 3x²y² + 2y, and ∂f/∂y = 2x³y + 2x. Pure second partials: ∂²f/∂x² = ∂/∂x(3x²y² + 2y) = 6xy². ∂²f/∂y² = ∂/∂y(2x³y + 2x) = 2x³. Mixed partial (differentiate x-partial with respect to y): ∂²f/∂x∂y = ∂/∂y(3x²y² + 2y) = 6x²y + 2. Mixed partial (differentiate y-partial with respect to x): ∂²f/∂y∂x = ∂/∂x(2x³y + 2x) = 6x²y + 2.
2. Clairaut's Theorem — why the mixed partials matched
Both mixed partials came out to 6x²y + 2, and this is not a coincidence. Clairaut's Theorem states that if ∂²f/∂x∂y and ∂²f/∂y∂x are both continuous near a point, they are equal at that point. For every polynomial, exponential, and trigonometric function you will encounter in a standard calculus course, the mixed partials always match — computing both is a built-in check on your work. If your two mixed partials disagree, you made an algebra error in one of them; go back and recheck. ✓ 6x²y + 2 = 6x²y + 2, confirming both computations are correct.
Clairaut's Theorem: for any function with continuous second partial derivatives, ∂²f/∂x∂y = ∂²f/∂y∂x. Use this as a free accuracy check every time you compute mixed partials.
Common Mistakes When Computing Partial Derivatives
These errors account for the majority of lost points on multivariable calculus exams. A partial derivative calculator step by step catches them automatically, but recognizing them yourself is what builds real understanding.
1. Forgetting to treat the other variable as a constant
The single most common mistake: differentiating x²y with respect to x as if y were also a variable to eliminate, producing 2x instead of the correct 2xy. Every letter that is not the active variable stays exactly as written, multiplied through like any other constant.
2. Dropping a term that has no active variable — instead of setting it to zero
For f(x, y) = x²y + y³, when finding ∂f/∂x, the term y³ has no x in it, so its derivative with respect to x is 0 — it disappears from the answer, but only because its derivative is genuinely zero, not because you skipped it. Writing ∂f/∂x = 2xy + 3y² (forgetting to zero out y³) is a frequent slip.
3. Mixing up which variable is 'active' mid-problem
Switching from differentiating with respect to x to accidentally treating y as active partway through a long expression produces an answer that satisfies neither partial derivative. Label your work clearly: write '∂/∂x, y held constant' above the line before you start.
4. Applying the Product Rule when it isn't needed
In x²y³, both factors depend on the same active variable only if you are differentiating with respect to a variable in both — but with respect to x, y³ is a constant coefficient, so ∂/∂x(x²y³) = 2xy³ directly (Power Rule, not Product Rule). The Product Rule for partials is only needed when two functions of the active variable are multiplied together, like x²·sin(x) — not when one factor is entirely a frozen constant.
5. Sign errors in mixed second partials
When a term includes a negative sign, such as −3xy², carrying that sign through two rounds of differentiation is where most mixed-partial mistakes happen. Compute ∂²f/∂x∂y and ∂²f/∂y∂x separately and compare — by Clairaut's Theorem they must match, so a mismatch means you dropped or flipped a sign somewhere.
Practice Problems: Partial Derivatives with Full Solutions
Attempt each problem before checking the solution. Problems progress from basic power-rule partials to a three-variable function and a mixed exponential-logarithm challenge. Use a partial derivative calculator to verify each answer after you try it yourself. Problem 1 — Power Rule: Find ∂f/∂x and ∂f/∂y for f(x, y) = 5x²y³. ∂f/∂x = 10xy³ (differentiate x², keep y³ as a constant multiplier) ∂f/∂y = 15x²y² (differentiate y³, keep 5x² as a constant multiplier) ✓ Problem 2 — Trigonometric Chain Rule: Find ∂f/∂x and ∂f/∂y for f(x, y) = sin(xy). Treat xy as the inner function u. ∂u/∂x = y, so ∂f/∂x = y·cos(xy). ∂u/∂y = x, so ∂f/∂y = x·cos(xy) ✓ Problem 3 — Three variables: Find ∂f/∂x, ∂f/∂y, and ∂f/∂z for f(x, y, z) = x²y + yz³ − 3xz. ∂f/∂x = 2xy − 3z (y term keeps y, z term vanishes, −3xz differentiates to −3z) ∂f/∂y = x² + z³ (x²y differentiates to x², yz³ differentiates to z³, −3xz vanishes) ∂f/∂z = 3yz² − 3x (x²y vanishes, yz³ differentiates to 3yz², −3xz differentiates to −3x) ✓ Problem 4 — Quotient written as a power: Find ∂f/∂x and ∂f/∂y for f(x, y) = x²/y = x²·y⁻¹. ∂f/∂x = 2x/y (Power Rule on x², y⁻¹ is a constant multiplier) ∂f/∂y = x²·(−1)y⁻² = −x²/y² ✓ Problem 5 — Challenge, exponential plus logarithm: Find ∂f/∂x and ∂f/∂y for f(x, y) = e^(xy) + x·ln(y), then evaluate both at (1, 1). ∂f/∂x = y·e^(xy) + ln(y) (Chain Rule on e^(xy) gives y·e^(xy); Product Rule on x·ln(y) with respect to x gives ln(y)) ∂f/∂y = x·e^(xy) + x/y (Chain Rule on e^(xy) gives x·e^(xy); Product Rule on x·ln(y) with respect to y gives x·(1/y)) At (1, 1): ∂f/∂x = 1·e¹ + ln(1) = e + 0 ≈ 2.718 ∂f/∂y = 1·e¹ + 1/1 = e + 1 ≈ 3.718 ✓
Why Do Partial Derivatives Matter in Real Applications?
Partial derivatives are not just an exam topic — they are the working tool behind entire fields. In physics, the heat equation and the wave equation are both written using second-order partial derivatives with respect to position and time; ∂T/∂t describes how temperature at a point changes over time, while ∂²T/∂x² describes how it curves across space. In economics, if profit π depends on price p and advertising spend a, ∂π/∂p (marginal profit from price changes) and ∂π/∂a (marginal profit from advertising) tell a firm which lever to adjust first. The gradient — the vector of all first partial derivatives, ∇f = (∂f/∂x, ∂f/∂y) — points in the direction of steepest increase of a function and has magnitude equal to that steepest slope. This is exactly what gradient descent uses in machine learning: a model with thousands of parameters computes the partial derivative of its loss function with respect to every parameter, then adjusts each one slightly against its gradient to reduce error. Every time a neural network 'learns,' it is really just computing a very large number of partial derivatives, all at once, over and over. Optimization problems with multiple constraints — maximizing area given a fixed perimeter, or minimizing cost given a production target — set every partial derivative to zero simultaneously to locate critical points, the same core idea as single-variable optimization extended to more than one direction at a time.
The gradient ∇f = (∂f/∂x, ∂f/∂y) is a vector built entirely from partial derivatives, and it points in the direction where f increases fastest — the mathematical basis of gradient descent in machine learning.
What Are Common Questions About Partial Derivative Calculators?
1. How is a partial derivative different from a normal derivative?
An ordinary derivative d/dx applies to a function of one variable, f(x). A partial derivative ∂f/∂x applies to a function of two or more variables, f(x, y, ...), and requires you to freeze every variable except x while differentiating. The mechanics of differentiation are identical — only the number of variables and the 'hold everything else constant' rule are new.
2. Do I always need to compute every partial derivative, or just one?
It depends on the question. Finding the gradient or checking for critical points requires every first partial derivative. Finding the rate of change in one specific direction, like 'how fast does temperature change as x increases with y fixed,' only requires that single partial derivative. Read the problem carefully to see which variable is asked about.
3. Why do the two mixed partial derivatives always come out equal?
This is Clairaut's Theorem: for any function whose second partial derivatives are continuous, ∂²f/∂x∂y equals ∂²f/∂y∂x, regardless of which order you differentiate in. Nearly every function in an introductory calculus course meets this continuity condition, so treat a mismatch between your two mixed partials as a signal to recheck your algebra.
4. Can a function have a partial derivative at a point but not be differentiable there overall?
Yes. A function can have well-defined partial derivatives ∂f/∂x and ∂f/∂y at a point without being fully (jointly) differentiable there, because full differentiability requires the function to be well-approximated by a single linear plane in every direction — not just along the x-axis and y-axis. This distinction matters in more advanced multivariable analysis courses but rarely affects standard homework problems built from polynomials, exponentials, and trig functions.
5. What does a step-by-step partial derivative calculator show that a plain answer doesn't?
A good partial derivative calculator step by step shows which variable was held constant at each stage, which differentiation rule applied to each term, and how each term simplified — the same reasoning an instructor expects on a graded solution. Comparing your own work line by line against that breakdown is the fastest way to find exactly where a mistake happened, rather than only learning that the final answer was wrong.
Related Articles
Derivative Calculator Step by Step: Complete Guide with Worked Examples
Master single-variable differentiation rules first — the foundation every partial derivative calculation builds on.
Calculus Help: Derivatives, Integrals, and Limits Explained
A full overview of calculus concepts, useful context beyond a single partial derivative problem.
Step-by-Step Math Problem Solving Guide
General strategies for breaking down multi-part math problems, including multivariable calculus.
Related Math Solvers
Step-by-Step Solutions
Get detailed explanations for every step, not just the final answer.
AI Math Tutor
Ask follow-up questions and get personalized explanations 24/7.
Smart Scan Solver
Snap a photo of any math problem and get an instant step-by-step solution.