Eigenvalue and Eigenvector Calculator: Step-by-Step Guide
An eigenvalue and eigenvector calculator finds the scalars λ and nonzero vectors v that satisfy Av = λv for a square matrix A, using the characteristic equation det(A - λI) = 0 to solve for eigenvalues and then back-substituting to find each eigenvector. Eigenvalues and eigenvectors reveal how a linear transformation stretches, shrinks, or reflects space along specific directions, and they underpin techniques from principal component analysis to structural vibration analysis and Google's PageRank algorithm. This guide walks through the full method step by step with real arithmetic, covering both a general 2×2 matrix and the diagonal/triangular shortcut, then works through practice problems with complete solutions.
Contents
- 01What Are Eigenvalues and Eigenvectors?
- 02How Does an Eigenvalue and Eigenvector Calculator Find Them Step by Step?
- 03Worked Example: Finding Eigenvalues and Eigenvectors of a 2×2 Matrix
- 04How Do You Find Eigenvalues of a Diagonal or Triangular Matrix?
- 05What Do Eigenvalues and Eigenvectors Mean Geometrically?
- 06Where Do Eigenvalues and Eigenvectors Show Up in Real Applications?
- 07Common Mistakes When Computing Eigenvalues and Eigenvectors
- 08Practice Problems: Test Your Understanding of Eigenvalues and Eigenvectors
- 09Why Use an Eigenvalue and Eigenvector Calculator for Step-by-Step Practice?
What Are Eigenvalues and Eigenvectors?
An eigenvalue and eigenvector calculator is built around one core equation: for a square n×n matrix A, a nonzero vector v is called an eigenvector if multiplying A by v produces a vector that points in the same direction as v, just scaled by a constant. That constant, written λ (lambda), is the eigenvalue. In equation form, Av = λv, where v ≠ 0. Rearranging gives (A - λI)v = 0, where I is the identity matrix of the same size as A. This is a homogeneous linear system, and it only has a nonzero solution v when the matrix (A - λI) is singular — meaning its determinant is zero. That single requirement, det(A - λI) = 0, is called the characteristic equation, and solving it is the first step to finding every eigenvalue of A. When a matrix A has n linearly independent eigenvectors, those eigenvectors can be assembled as the columns of a matrix P, and A can be rewritten as A = PDP⁻¹, where D is the diagonal matrix holding the corresponding eigenvalues. This process is called diagonalization, and it's the main reason eigenvalues and eigenvectors are studied together rather than separately — an 'eigendecomposition' of A is only possible once you have both pieces.
Av = λv: an eigenvector v keeps its direction under the transformation A — it is only stretched, shrunk, or flipped by the scalar eigenvalue λ.
How Does an Eigenvalue and Eigenvector Calculator Find Them Step by Step?
A step-by-step eigenvalue and eigenvector calculator always follows the same four-stage process, whether the matrix is 2×2 or 6×6. The only thing that changes with size is how much arithmetic each stage requires. Not every matrix behaves this cleanly. A repeated eigenvalue can come with either a full set of independent eigenvectors or fewer than expected — a situation called a defective matrix — and a non-symmetric real matrix can produce a pair of complex conjugate eigenvalues with complex eigenvectors. An eigenvalue and eigenvector calculator handles all three cases with the same characteristic-equation method; only the interpretation of the result changes once you reach Step 2.
1. Step 1 — Build the characteristic equation
Subtract λ from each diagonal entry of A to form (A - λI), then set its determinant equal to zero: det(A - λI) = 0. Expanding this determinant produces a polynomial in λ called the characteristic polynomial.
2. Step 2 — Solve the characteristic polynomial for λ
Factor or use the quadratic formula (for 2×2 matrices) or polynomial factoring techniques (for larger matrices) to find every root. Each root is one eigenvalue of A; an n×n matrix has exactly n eigenvalues, counted with multiplicity, possibly including repeats or complex numbers.
3. Step 3 — Substitute each eigenvalue back to find its eigenvector
For each eigenvalue λ, plug it into (A - λI)v = 0 and solve the resulting linear system for v. Because det(A - λI) = 0 by construction, this system always has infinitely many solutions — pick the simplest nonzero vector that satisfies it.
4. Step 4 — Verify with Av = λv
Multiply the original matrix A by the eigenvector v you found and confirm the result equals λ times v. This check catches arithmetic slips before they propagate into an incorrect final answer.
5. Step 5 — Handle larger matrices with cofactor expansion
For 3×3 and larger matrices, det(A - λI) is expanded using cofactor (minor) expansion along a row or column, which turns the determinant into a sum of signed smaller determinants that each contain λ. The result is a degree-n polynomial in λ. When no diagonal or triangular shortcut applies, look for rational roots first using the constant term of the expanded polynomial, then factor out each root you find before solving whatever quadratic or linear factor remains.
Every eigenvalue problem reduces to two steps: solve det(A - λI) = 0 for λ, then solve (A - λI)v = 0 for v.
Worked Example: Finding Eigenvalues and Eigenvectors of a 2×2 Matrix
Consider A = [[4, 1], [2, 3]]. This is a standard 2×2 matrix with two distinct real eigenvalues, making it a clean first example of the full method.
1. Step 1 — Set up and expand the characteristic equation
(A - λI) = [[4-λ, 1], [2, 3-λ]] det(A - λI) = (4-λ)(3-λ) - (1)(2) = λ² - 7λ + 12 - 2 = λ² - 7λ + 10
2. Step 2 — Factor and solve for the eigenvalues
λ² - 7λ + 10 = (λ - 5)(λ - 2) = 0, so λ₁ = 5 and λ₂ = 2.
3. Step 3 — Find the eigenvector for λ₁ = 5
(A - 5I) = [[-1, 1], [2, -2]] Row 1 gives -x + y = 0, so y = x. Choosing x = 1 gives eigenvector v₁ = (1, 1).
4. Step 4 — Find the eigenvector for λ₂ = 2
(A - 2I) = [[2, 1], [2, 1]] Row 1 gives 2x + y = 0, so y = -2x. Choosing x = 1 gives eigenvector v₂ = (1, -2).
5. Step 5 — Verify both results
Av₁ = [4(1)+1(1), 2(1)+3(1)] = [5, 5] = 5(1, 1) ✓ Av₂ = [4(1)+1(-2), 2(1)+3(-2)] = [2, -4] = 2(1, -2) ✓ Both eigenpairs check out.
6. Step 6 — Confirm the matrix is diagonalizable
Because both eigenvalues are real and distinct, A is diagonalizable: assembling v₁ and v₂ as the columns of P = [[1, 1], [1, -2]] and D = [[5, 0], [0, 2]] recovers the original matrix through A = PDP⁻¹. Every 2×2 matrix with two distinct eigenvalues is guaranteed to be diagonalizable this way.
7. Step 7 — Cross-check using trace and determinant
Two quick sanity checks work for any square matrix: the sum of the eigenvalues always equals the trace of A (the sum of its diagonal entries), and the product of the eigenvalues always equals det(A). Here, 5 + 2 = 7 matches trace(A) = 4 + 3 = 7, and 5 × 2 = 10 matches det(A) = (4)(3) - (1)(2) = 10 — confirming both eigenvalues are correct without re-deriving them.
Result: A = [[4,1],[2,3]] has eigenvalues λ₁ = 5 with eigenvector (1, 1), and λ₂ = 2 with eigenvector (1, -2).
How Do You Find Eigenvalues of a Diagonal or Triangular Matrix?
Diagonal and triangular matrices offer a shortcut that skips the characteristic-equation arithmetic entirely: their eigenvalues are simply the entries sitting on the main diagonal. This works because subtracting λ from the diagonal of a triangular matrix keeps it triangular, and the determinant of any triangular matrix is just the product of its diagonal entries.
1. Step 1 — Read eigenvalues directly from a diagonal matrix
For D = [[3, 0, 0], [0, -1, 0], [0, 0, 5]], det(D - λI) = (3-λ)(-1-λ)(5-λ), so the eigenvalues are simply λ = 3, λ = -1, and λ = 5 — no elimination required. Their eigenvectors are the standard basis vectors (1,0,0), (0,1,0), and (0,0,1).
2. Step 2 — Apply the same shortcut to an upper triangular matrix
For T = [[2, 4, 1], [0, 5, 3], [0, 0, -3]], det(T - λI) = (2-λ)(5-λ)(-3-λ) because the matrix stays upper triangular after subtracting λ. The eigenvalues are the diagonal entries: λ = 2, λ = 5, and λ = -3.
3. Step 3 — Find an eigenvector for a triangular matrix using back-substitution
For λ = 5 in T: (T - 5I) = [[-3, 4, 1], [0, 0, 3], [0, 0, -8]]. Row 2 gives 3z = 0, so z = 0. Row 1 gives -3x + 4y + z = 0, so -3x + 4y = 0. Choosing y = 3 gives x = 4, so v = (4, 3, 0).
4. Step 4 — Verify the triangular eigenvector
Tv = [2(4)+4(3)+1(0), 5(3), -3(0)] = [20, 15, 0] = 5(4, 3, 0) ✓, confirming λ = 5 pairs correctly with eigenvector (4, 3, 0).
5. Step 5 — Understand why the shortcut matters beyond speed
This shortcut is more than a convenience — it's also the reason diagonalization is useful in the first place. If you can find a matrix P that converts A into a triangular or diagonal matrix D, computing large powers of A (needed for solving recurrence relations and systems of differential equations) collapses into computing powers of D, which just means raising each diagonal entry to that power.
Shortcut: for any diagonal or triangular matrix, the eigenvalues are exactly the diagonal entries — det(A - λI) factors automatically into (d₁-λ)(d₂-λ)⋯(dₙ-λ).
What Do Eigenvalues and Eigenvectors Mean Geometrically?
Picture a matrix A as a transformation that moves every point in space. Most vectors get pushed off in a new direction when you apply A, but eigenvectors are the special directions that A leaves unrotated — a vector along an eigenvector direction only gets stretched, compressed, or flipped, never turned. The eigenvalue λ tells you exactly how much: |λ| > 1 stretches the eigenvector direction, 0 < |λ| < 1 compresses it toward the origin, and a negative λ flips it to point the opposite way while still lying on the same line. In the 2×2 example above, every vector along the line through the origin and (1, 1) gets stretched by a factor of 5 under A, while every vector along the line through (1, -2) gets stretched by a factor of 2. Any other vector in the plane is a combination of these two directions, so its path under repeated application of A is dominated more and more by the larger eigenvalue — the same idea that powers Markov chain convergence and iterative algorithms like PageRank. Not every matrix has enough independent eigenvector directions to span the whole space. When it doesn't, no diagonal matrix D can represent it exactly, and mathematicians instead work with a related structure called the Jordan form. This is the geometric reason repeated eigenvalues sometimes cause trouble: two directions can collapse into a single axis, leaving fewer than n independent directions for an n×n matrix to diagonalize along.
1. Quick example — a reflection matrix
For A = [[0, 1], [1, 0]], solving det(A - λI) = λ² - 1 = 0 gives λ = 1 and λ = -1. The eigenvector for λ = 1 is (1, 1) — the line y = x, which A leaves completely untouched — while the eigenvector for λ = -1 is (1, -1), the line that A flips end to end. This is exactly how the matrix earns its geometric name: it reflects the plane across the line y = x, and the two eigenvalues (1 and -1) capture the 'keep' and 'flip' directions precisely.
Geometrically, eigenvectors are the axes a transformation respects; eigenvalues are how much it stretches, shrinks, or flips along each axis.
Where Do Eigenvalues and Eigenvectors Show Up in Real Applications?
Eigenvalues and eigenvectors are not just a linear algebra exercise — they are the mathematical backbone of several tools used daily in engineering, data science, and physics. In each case below, the same characteristic-equation method from the sections above is the starting point; only the matrix being analyzed changes.
1. Principal Component Analysis (PCA)
Data scientists compute the eigenvectors of a dataset's covariance matrix to find the directions of maximum variance, then use the largest eigenvalues to decide which directions (principal components) capture the most information for dimensionality reduction.
2. Structural and mechanical vibration analysis
Engineers model a structure's natural vibration frequencies as the eigenvalues of a stiffness/mass system, and the eigenvectors describe the corresponding mode shapes — how the structure physically bends at each frequency.
3. Search ranking algorithms
Google's original PageRank algorithm ranks web pages using the dominant eigenvector of a matrix built from the web's link structure, since that eigenvector represents the long-run steady-state distribution of a random surfer.
4. Stability of dynamical systems
In differential equations and control theory, the sign and magnitude of a system's eigenvalues determine whether a system returns to equilibrium, oscillates, or grows without bound over time.
Every one of these applications reduces to the same question this guide answers: which directions does a matrix respect, and by how much does it scale them?
Common Mistakes When Computing Eigenvalues and Eigenvectors
A handful of errors account for most incorrect eigenvalue and eigenvector calculations done by hand. Watch for these five in particular before you trust a final answer.
1. Mistake 1 — Subtracting λ from every entry instead of just the diagonal
Only the diagonal entries of A change when forming A - λI; the off-diagonal entries stay exactly as they are in the original matrix.
2. Mistake 2 — Treating an eigenvector as unique
Eigenvectors are only defined up to a nonzero scalar multiple — (1, 1), (2, 2), and (-3, -3) are all valid eigenvectors for the same eigenvalue. There is no single 'correct' eigenvector, only a correct direction.
3. Mistake 3 — Sign errors while expanding larger determinants
For 3×3 and larger matrices, cofactor expansion involves alternating signs; dropping or misplacing a minus sign is the single most common source of a wrong characteristic polynomial.
4. Mistake 4 — Applying the diagonal shortcut to a non-triangular matrix
The 'eigenvalues equal the diagonal entries' shortcut only holds for diagonal and triangular matrices. For a general matrix like [[4,1],[2,3]], the diagonal entries 4 and 3 are not the eigenvalues — you still need the full characteristic equation.
5. Mistake 5 — Assuming eigenvalues must be real
A non-symmetric matrix can have complex conjugate eigenvalues with complex-entry eigenvectors, even when every entry of the original matrix is a real number. This shows up most often with rotation-like matrices, so don't discard a result just because the characteristic polynomial has no real roots — solve it over the complex numbers instead.
The fastest way to catch an eigenvector mistake is to recompute Av and check whether it truly equals λv before moving on.
Practice Problems: Test Your Understanding of Eigenvalues and Eigenvectors
Work through these three problems, then compare your steps to the answers below.
1. Problem 1 — Find the eigenvalues and eigenvectors of A = [[1, 2], [2, 1]]
Answer: det(A-λI) = (1-λ)² - 4 = λ² - 2λ - 3 = (λ-3)(λ+1) = 0, so λ₁ = 3 and λ₂ = -1. For λ₁ = 3: (A-3I) = [[-2,2],[2,-2]] gives x = y, so v₁ = (1,1). For λ₂ = -1: (A+I) = [[2,2],[2,2]] gives x = -y, so v₂ = (1,-1).
2. Problem 2 — Find the eigenvalues of the upper triangular matrix U = [[6, 2], [0, -2]]
Answer: U is triangular, so its eigenvalues are read directly from the diagonal: λ₁ = 6 and λ₂ = -2. No characteristic-equation arithmetic needed.
3. Problem 3 — Is v = (2, 1) an eigenvector of A = [[3, 4], [0, 1]]?
Answer: Av = [3(2)+4(1), 0(2)+1(1)] = (10, 1). For v to be an eigenvector, (10, 1) would need to be a scalar multiple of (2, 1), but 10/2 = 5 while 1/1 = 1 — the ratios don't match, so v = (2, 1) is not an eigenvector of A.
4. Problem 4 — Find the eigenvalues of the upper triangular matrix M = [[2, 1, 0], [0, 2, 5], [0, 0, 7]]
Answer: M is upper triangular, so its eigenvalues are the diagonal entries: λ = 2, appearing with multiplicity 2 since it occupies two diagonal positions, and λ = 7. A repeated eigenvalue like this one is exactly the case where you'd need to check separately whether two independent eigenvectors exist for λ = 2 or only one — a distinction that determines whether M is diagonalizable.
Checking whether a given vector is an eigenvector is simple: compute Av and see if every component scales by the same constant.
Why Use an Eigenvalue and Eigenvector Calculator for Step-by-Step Practice?
Computing eigenvalues by hand is quick for a 2×2 matrix but the arithmetic grows fast for 3×3 and larger systems, where a single sign error in the characteristic polynomial throws off every eigenvalue and eigenvector that follows. An eigenvalue and eigenvector calculator that shows each stage — the characteristic equation, the factored polynomial, and the back-substitution for every eigenvector — lets you check your own work line by line instead of only seeing whether the final answer matches. Solvify's step-by-step solver and AI math tutor walk through characteristic equations, determinants, and eigenvector back-substitution the same way this guide does, so you can verify homework, prepare for exams, or get unstuck on a specific matrix without losing the reasoning behind each step. This is especially useful partway through a linear algebra course, where eigenvalue problems build directly on the matrix operations, determinants, and row reduction covered in the related guides below.
Related Articles
Matrix Rank Calculator: How to Find Rank with Row Echelon Form
Rank and eigenvalues both come from the same matrix, and a singular (A - λI) always has rank less than n — see the row-reduction method in full detail.
Gaussian Elimination Calculator: Solve Linear Systems Step by Step
Eigenvector back-substitution is really just solving the homogeneous system (A - λI)v = 0 — this guide covers the elimination method behind every step.
Matrix Calculator Step by Step: Operations, Determinants, and Inverses
Determinants are the core tool behind the characteristic equation det(A - λI) = 0 — this guide covers determinants alongside the rest of core matrix operations.
Related Math Solvers
Step-by-Step Solutions
Get detailed explanations for every step, not just the final answer.
Smart Scan Solver
Snap a photo of any math problem and get an instant step-by-step solution.
AI Math Tutor
Ask follow-up questions and get personalized explanations 24/7.