8.3 Transformations

Transformations are mathematical operations that change geometric objects in specific, predictable ways. They give you a framework for understanding spatial relationships, from simple slides and flips to complex compositions that power computer graphics and physics simulations. This guide covers the major types of transformations, how they work in coordinate systems, and where they connect to broader mathematical structures.

Types of transformations

Transformations alter geometric objects according to defined rules. Categorizing them helps you predict what properties a shape will keep (or lose) after the transformation is applied.

Rigid vs non-rigid transformations

Rigid transformations (also called isometries) preserve distances and angles between points. The shape and size of the object stay exactly the same. Translations, rotations, and reflections are all rigid.

Non-rigid transformations change the shape or size of an object. Dilations (scaling) and shears fall into this category. A dilation can produce a similar figure (same shape, different size), while a shear can distort the shape entirely.

Linear vs non-linear transformations

A linear transformation preserves two operations: vector addition and scalar multiplication. In coordinate systems, you can represent any linear transformation as matrix multiplication. Rotations, reflections, and scaling are all linear.

Non-linear transformations can bend straight lines into curves or produce other distortions. Exponential and logarithmic mappings are common examples. If the output can't be captured by multiplying a matrix times a vector, the transformation is non-linear.

Affine transformations

Affine transformations combine a linear transformation with a translation. They preserve collinearity (points on a line stay on a line) and ratios of distances along a line. Shears, scaling, rotations, translations, and any combination of these are affine.

To represent affine transformations as a single matrix operation, you use homogeneous coordinates or augmented matrices. This trick turns the translation component into part of the matrix multiplication.

Translation

A translation slides every point of a figure the same distance in the same direction. Nothing about the shape changes: no rotation, no resizing, no distortion.

Vector representation

Translations are described by a displacement vector that specifies how far and in what direction to move. In 2D, this vector is an ordered pair $(a, b)$, where $a$ is the horizontal shift and $b$ is the vertical shift. You perform the translation by adding this vector to each point's coordinates.

Coordinate changes

For a translation by vector $(a, b)$:

$(x, y) \rightarrow (x + a, y + b)$

For example, translating the point $(3, 5)$ by the vector $(2, -1)$ gives $(5, 4)$. Translations preserve distances, angles, and orientation. The figure doesn't flip or turn at all.

Applications in geometry

• Solving problems involving parallel lines or congruent figures (slide one figure onto another to show congruence)
• Describing physical displacement in physics
• Moving objects on screen in computer graphics
• Analyzing translational symmetry in repeating patterns like wallpaper or tiling

Rotation

A rotation turns every point of a figure around a fixed point by a specified angle. Rotations preserve shape and size, making them rigid transformations. They also connect directly to trigonometry and complex number multiplication.

Angle of rotation

The angle is measured in degrees or radians. By convention, positive angles indicate counterclockwise rotation and negative angles indicate clockwise rotation. A full rotation is $360°$ or $2\pi$ radians.

Center of rotation

The center of rotation is the fixed point that doesn't move. The origin $(0, 0)$ is the most common choice because it simplifies the math. If you need to rotate around a different point, you:

1. Translate the figure so the center of rotation moves to the origin
2. Apply the rotation
3. Translate back

Rotation matrices

The 2D rotation matrix for angle $\theta$ is:

$\begin{bmatrix} \cos \theta & -\sin \theta \\ \sin \theta & \cos \theta \end{bmatrix}$

To rotate a point, multiply this matrix by the point's coordinate vector. Two useful properties:

• Composing rotations: multiply the rotation matrices together (a rotation by $\alpha$ followed by $\beta$ equals a rotation by $\alpha + \beta$)
• Inverse rotation: transpose the rotation matrix (equivalently, negate the angle)

Reflection

A reflection flips a figure across a line (or point), producing a mirror image. Reflections are rigid transformations, but unlike translations and rotations, they reverse orientation (a clockwise-labeled shape becomes counterclockwise).

Line of reflection

The line of reflection acts as a mirror. Every point maps to a new position such that the line is the perpendicular bisector of the segment connecting the original and reflected points.

Some common reflection lines in 2D:

• Reflection across the x-axis: $(x, y) \rightarrow (x, -y)$
• Reflection across the y-axis: $(x, y) \rightarrow (-x, y)$
• Reflection across the line $y = x$: $(x, y) \rightarrow (y, x)$

Point of reflection

A point reflection maps each point through a central point to the opposite side, at equal distance. The center is the midpoint between each original point and its image.

A point reflection through the origin gives $(x, y) \rightarrow (-x, -y)$, which is equivalent to a $180°$ rotation.

Symmetry in transformations

• Reflectional symmetry: a figure looks the same on both sides of a mirror line
• Rotational symmetry: a figure looks identical after rotation by some angle less than $360°$
• Translational symmetry: a pattern repeats at regular intervals

Symmetry groups classify figures by which transformations leave them unchanged. A square, for instance, has a symmetry group with 8 elements (4 rotations and 4 reflections).

Scaling

Scaling changes the size of a figure by multiplying coordinates by a scale factor. When the scale factor is the same in every direction, the shape is preserved. Scaling is central to proportional reasoning, map-making, and modeling.

Uniform vs non-uniform scaling

Uniform scaling applies the same factor to all dimensions, so the figure grows or shrinks proportionally. Shape and angles are preserved.

Non-uniform scaling uses different factors for different dimensions. This distorts the figure: a circle becomes an ellipse, a square becomes a rectangle. Angles generally change.

Scale factors

• A factor greater than 1 enlarges the object
• A factor between 0 and 1 shrinks it
• A factor of 1 leaves it unchanged
• A negative factor reflects the object (flips it) in addition to scaling
• Reciprocal factors reverse each other: scaling by 3 then by $\frac{1}{3}$ returns to the original size

Effects on shape properties

This is where students often trip up on exams. The relationships follow a pattern:

• Linear measurements (perimeter, side lengths) scale by the factor $k$
• Area scales by $k^2$
• Volume scales by $k^3$
• Angles remain unchanged under uniform scaling

So if you double the side length of a square ($k = 2$), its area quadruples ($2^2 = 4$).

Shear transformations

A shear slides layers of a figure relative to each other, like pushing the top of a deck of cards sideways. Shears distort shape but, notably, preserve area.

Horizontal vs vertical shear

A horizontal shear shifts points parallel to the x-axis. The farther a point is from the x-axis, the more it shifts. A vertical shear does the same but parallel to the y-axis. The shear factor $k$ controls how much distortion occurs.

Shear matrices

Horizontal shear:

$\begin{bmatrix} 1 & k \\ 0 & 1 \end{bmatrix}$

Vertical shear:

$\begin{bmatrix} 1 & 0 \\ k & 1 \end{bmatrix}$

Here $k$ is the shear factor. You can combine shear matrices with other transformation matrices through multiplication.

Applications in engineering

• Analyzing material deformation under stress (shear stress in beams)
• Modeling laminar flow in fluid dynamics
• Correcting skew in scanned documents and image processing
• Designing structures to resist shear forces

Composition of transformations

You can combine multiple transformations into a single operation. This is called composition, and it's the foundation of complex geometric manipulations in graphics, animation, and physics.

Order of transformations

Order matters. Applying transformation A then B generally gives a different result than B then A. In mathematical terms, composition is non-commutative: $AB \neq BA$ in general.

For example, rotating a shape and then translating it produces a different result than translating first and then rotating.

Matrix multiplication

To compose transformations represented as matrices:

1. Write each transformation as a matrix
2. Multiply the matrices together, but in reverse order of application (if you apply A first, then B, the composite matrix is $BA$)
3. The resulting matrix represents the entire composite transformation

The identity matrix acts as a "do nothing" transformation: multiplying by it leaves any transformation unchanged.

Inverse transformations

An inverse transformation undoes the effect of the original. If transformation $T$ maps point $P$ to $P'$, then $T^{-1}$ maps $P'$ back to $P$.

For a composite transformation $AB$, the inverse is $B^{-1}A^{-1}$ (reverse the order and invert each piece). Not every transformation has an inverse: if the transformation matrix is singular (determinant equals zero), the transformation collapses dimensions and can't be reversed.

Transformations in coordinate systems

Different coordinate systems make certain transformations simpler. Choosing the right system for the problem can save significant work.

Cartesian coordinates

The standard $(x, y)$ system. Linear transformations are represented by $2 \times 2$ matrices, and translations require adding a vector separately. This system works best for problems involving straight lines and rectangular grids.

Polar coordinates

Points are represented by $(r, \theta)$: distance from the origin and angle from the positive x-axis. Rotations become simple addition: rotating by angle $\alpha$ just means $\theta \rightarrow \theta + \alpha$. Scaling affects only $r$. Polar coordinates are natural for circular and spiral patterns.

Homogeneous coordinates

Homogeneous coordinates represent a 2D point as $(x, y, w)$, where the Cartesian equivalent is $(x/w, y/w)$. The key advantage: translations, rotations, scaling, and other affine transformations can all be expressed as matrix multiplication in this system. This unification is why homogeneous coordinates are standard in computer graphics pipelines.

They also allow representation of points at infinity, which matters in projective geometry.

Transformations in higher dimensions

The same transformation concepts extend beyond 2D. Working in higher dimensions builds abstract reasoning skills and connects to physics and data science.

3D transformations

3D transformations extend 2D ideas with some added complexity. Rotation in 3D happens around an axis rather than a point, and rotation matrices become $3 \times 3$. Three independent rotations are often described as pitch (tilting up/down), yaw (turning left/right), and roll (tilting side to side).

Hyperspace transformations

Transformations in spaces with more than three dimensions generalize rotation, reflection, and scaling. These are important in theoretical physics (e.g., spacetime in relativity) and in machine learning (transforming high-dimensional feature spaces). Visualization relies on techniques like projection and cross-sectional slicing.

Projective geometry

Projective geometry studies properties that remain unchanged under projective transformations (which include all affine transformations plus perspective projections). It incorporates points at infinity, where parallel lines "meet." Homogeneous coordinates represent projective transformations, and applications include computer vision, camera calibration, and 3D graphics rendering.

Applications of transformations

Computer graphics

• Rendering 3D scenes onto 2D displays using projection transformations
• Animating objects through sequences of composed transformations
• Image processing effects (blur, warp, color correction)
• Real-time rendering pipelines optimize transformation calculations using GPU hardware

Physics and engineering

• Describing motion and forces in classical mechanics (coordinate transformations between reference frames)
• Analyzing symmetries in particle physics (symmetry groups predict conservation laws)
• Modeling structural deformations in engineering
• Coordinate transformations in special and general relativity (Lorentz transformations)

Data visualization

• Projecting high-dimensional data onto 2D or 3D for visualization (e.g., PCA reduces dimensions by finding optimal rotation/projection)
• Scaling and normalizing data so different variables are comparable
• Rotating data views to reveal hidden patterns or clusters
• Interactive visualizations that let users apply transformations in real time

Properties of transformations

Understanding what a transformation preserves tells you what category it belongs to and what you can conclude about the transformed figure.

Preservation of angles

Conformal transformations preserve angles between curves at their intersection points. Rigid transformations and uniform scaling are conformal. This property is critical in cartography (conformal map projections preserve local shapes) and complex analysis (analytic functions are conformal).

Preservation of distances

Isometries preserve distances between all pairs of points. The isometries in 2D are translations, rotations, reflections, and glide reflections (a reflection followed by a translation along the reflection line). Two figures related by an isometry are congruent.

Preservation of parallelism

Affine transformations preserve parallel lines. If two lines are parallel before the transformation, they stay parallel after. This includes all linear transformations plus translations. Projective transformations, by contrast, can map parallel lines to lines that converge at a vanishing point.

Transformations in abstract algebra

Transformations connect geometry to algebra. Studying them through algebraic structures reveals deep patterns about symmetry and solvability.

Group theory

A set of transformations forms a group under composition if it satisfies four properties: closure, associativity, existence of an identity element, and existence of inverses. The set of all rotations about a point, for example, forms a group. Group structure captures the essential symmetry of a system.

Symmetry groups

Symmetry groups classify geometric objects by the transformations that leave them unchanged.

• Cyclic groups ($C_n$) describe pure rotational symmetry of regular polygons (e.g., $C_3$ for an equilateral triangle)
• Dihedral groups ($D_n$) include both rotations and reflections (e.g., $D_4$ for a square, with 8 elements)
• Continuous groups like $SO(3)$ describe rotational symmetry of 3D objects like spheres

These groups have direct applications in crystallography and molecular chemistry.

Galois theory

Galois theory studies symmetries of polynomial equations by examining automorphisms (structure-preserving transformations) of field extensions. It connects group theory to field theory and answers a fundamental question: which polynomial equations can be solved by radicals (using only addition, subtraction, multiplication, division, and roots)? The symmetry group of the equation's roots determines the answer.