Cylindrical and Spherical Coordinates

Learning Objectives

By the end of this section, you will be able to:

1. Understand the geometric meaning of cylindrical coordinates $(r, \theta, z)$ and when they simplify problems
2. Understand the geometric meaning of spherical coordinates $(\rho, \theta, \phi)$ and their relationship to latitude/longitude
3. Convert between Cartesian, cylindrical, and spherical coordinate systems
4. Derive the volume elements (Jacobians) for cylindrical and spherical coordinates
5. Apply these coordinate systems to simplify integration problems
6. Recognize applications in physics, engineering, and machine learning

The Big Picture: Beyond Cartesian Coordinates

"The art of choosing the right coordinate system is the art of making the impossible tractable." — Mathematical Wisdom

Cartesian coordinates $(x, y, z)$ are powerful and intuitive, but they are not always the best choice. Many natural objects and physical phenomena exhibit symmetry that Cartesian coordinates cannot easily capture:

• Cylinders, pipes, and towers have circular cross-sections and vertical axes
• Spheres, planets, and atoms are naturally described by distance from a center
• Rotation problems involve angles rather than linear displacements
• Wave propagation often spreads radially from a source

This section introduces two essential coordinate systems that complement Cartesian coordinates:

Historical Origins: The Evolution of Coordinate Systems

The development of non-Cartesian coordinate systems arose from practical needs in astronomy, physics, and engineering.

Ancient Astronomy and Celestial Coordinates

Long before René Descartes formalized Cartesian coordinates in 1637, astronomers used spherical-like systems to map the heavens. The celestial sphere — an imaginary sphere surrounding Earth — was described using:

• Declination: Angle above or below the celestial equator (like latitude)
• Right Ascension: Angle along the celestial equator (like longitude)

This is essentially a spherical coordinate system with Earth at the center.

Newton and Gravitational Fields (1687)

Isaac Newton's law of gravitation describes force as inversely proportional to the square of distance: $F \propto 1/r^2$. This radial dependence makes spherical coordinates natural for gravitational problems. Newton himself used geometric arguments that implicitly relied on spherical symmetry.

Euler and the Formalization (1750s)

Leonhard Euler systematically developed the mathematical framework for cylindrical and spherical coordinates, including:

• Explicit conversion formulas between coordinate systems
• Expressions for arc length, area, and volume elements
• Methods for transforming differential equations

Laplace and Potential Theory (1785)

Pierre-Simon Laplace expressed his famous equation $\nabla^2 \phi = 0$ in spherical coordinates, enabling solutions for gravitational and electrostatic potentials around spherical bodies. This became foundational for mathematical physics.

Cylindrical Coordinates

Cylindrical coordinates extend polar coordinates into three dimensions by adding a height coordinate $z$. A point is specified by three values:

Conversion Formulas: Cylindrical ↔ Cartesian

Geometric Interpretation

Think of cylindrical coordinates as describing a point by:

1. How far is it from the central axis (the z-axis)? This is $r$.
2. In what direction around the axis? This is $\theta$.
3. How high up or down? This is $z$.

The name "cylindrical" comes from the fact that surfaces of constant $r$ are cylinders centered on the z-axis.

Spherical Coordinates

Spherical coordinates describe a point by its distance from the origin and two angles. This system is ideal for problems with radial symmetry.

Conversion Formulas: Spherical ↔ Cartesian

Understanding the Angles

The two angles in spherical coordinates have distinct roles:

• $\theta$ (azimuthal angle): Like longitude on Earth, this angle sweeps around the z-axis in the xy-plane. It determines the "compass direction."
• $\phi$ (polar angle): Like colatitude (90° - latitude), this angle measures how far "down" from the north pole (positive z-axis) the point is.$\phi = 0$ is the north pole, $\phi = \pi/2$ is the equator, and $\phi = \pi$ is the south pole.

Conversion Formulas

Use this interactive calculator to convert between all three coordinate systems:

Cylindrical ↔ Spherical Conversions

Sometimes you need to convert directly between cylindrical and spherical coordinates:

Notice that $\theta$ is the same in both systems — both measure the angle in the xy-plane from the positive x-axis.

Volume Elements and Jacobians

When changing coordinates in an integral, we must account for how the coordinate transformation stretches or compresses space. This is captured by the Jacobian.

Cylindrical Coordinates: dV = r dr dθ dz

In cylindrical coordinates, the volume element is:

The factor of $r$ appears because:

• A small change $d\theta$ sweeps out an arc of length $r \cdot d\theta$
• Near the z-axis ($r \approx 0$), this arc is tiny
• Far from the z-axis, the same $d\theta$ covers more space

Spherical Coordinates: dV = ρ² sin(φ) dρ dθ dφ

In spherical coordinates, the volume element is:

This Jacobian has two factors:

• $\rho^2$: Spherical shells have area proportional to $\rho^2$ (surface area of sphere = $4\pi\rho^2$)
• $\sin\phi$: Circles of latitude shrink near the poles. At $\phi = 0$ (north pole), the circumference is zero; at $\phi = \pi/2$ (equator), it's maximum.

Real-World Applications

When to Use Each Coordinate System

Physics Applications

Many fundamental equations of physics simplify dramatically in the right coordinates:

• Electric field of a charged wire: Cylindrical coordinates, field depends only on $r$
• Gravitational field of a planet: Spherical coordinates, field depends only on $\rho$
• Schrödinger equation for hydrogen: Spherical coordinates allow separation of variables, yielding the famous atomic orbitals
• Fluid flow in a pipe: Cylindrical coordinates capture the parabolic velocity profile

Machine Learning Applications

Coordinate systems and their transformations appear throughout modern machine learning:

1. Hyperspherical Coordinates in High Dimensions

Spherical coordinates generalize to n dimensions as hyperspherical coordinates. These are crucial for:

• Uniform sampling on hyperspheres: Many ML algorithms need to sample uniformly from the surface of a hypersphere (e.g., for weight initialization)
• Directional statistics: When data lies on a sphere (text embeddings, normalized features), spherical coordinates provide natural distributions
• von Mises-Fisher distribution: The spherical analog of the Gaussian, used in clustering on spheres

2. Normalizing Flows and Coordinate Transforms

Normalizing flows transform simple distributions into complex ones using invertible mappings. The key insight is tracking the Jacobian:

This is exactly the change-of-variables formula from multivariable calculus! Polar/spherical transforms inspired many flow architectures.

3. 3D Computer Vision and Graphics

Modern 3D deep learning heavily uses coordinate systems:

• Point cloud processing: Converting between Cartesian and spherical representations for different tasks
• NeRF (Neural Radiance Fields): Ray marching uses parametric coordinates
• Spherical CNNs: Convolutions on spherical data (climate, planetary) require understanding spherical geometry

4. Positional Encodings

The sinusoidal positional encodings in Transformers:

These are essentially polar coordinates! The position is encoded as angles at different frequencies, allowing the model to learn relative positions through rotation-like operations.

Python Implementation

Coordinate Conversion Functions

1import numpy as np
2import matplotlib.pyplot as plt
3from mpl_toolkits.mplot3d import Axes3D
4
5def cartesian_to_cylindrical(x, y, z):
6    """
7    Convert Cartesian (x, y, z) to Cylindrical (r, theta, z).
8
9    r: radial distance from z-axis
10    theta: angle from positive x-axis in xy-plane
11    z: height (unchanged)
12    """
13    r = np.sqrt(x**2 + y**2)
14    theta = np.arctan2(y, x)  # Returns angle in [-pi, pi]
15    return r, theta, z
16
17def cylindrical_to_cartesian(r, theta, z):
18    """
19    Convert Cylindrical (r, theta, z) to Cartesian (x, y, z).
20    """
21    x = r * np.cos(theta)
22    y = r * np.sin(theta)
23    return x, y, z
24
25def cartesian_to_spherical(x, y, z):
26    """
27    Convert Cartesian (x, y, z) to Spherical (rho, theta, phi).
28
29    rho: distance from origin
30    theta: azimuthal angle (in xy-plane from x-axis)
31    phi: polar angle (from positive z-axis)
32    """
33    rho = np.sqrt(x**2 + y**2 + z**2)
34    theta = np.arctan2(y, x)
35    phi = np.arccos(z / rho) if rho > 0 else 0
36    return rho, theta, phi
37
38def spherical_to_cartesian(rho, theta, phi):
39    """
40    Convert Spherical (rho, theta, phi) to Cartesian (x, y, z).
41    """
42    x = rho * np.sin(phi) * np.cos(theta)
43    y = rho * np.sin(phi) * np.sin(theta)
44    z = rho * np.cos(phi)
45    return x, y, z
46
47# Example: Create a sphere using spherical coordinates
48rho = 2  # Radius
49theta = np.linspace(0, 2*np.pi, 50)
50phi = np.linspace(0, np.pi, 25)
51THETA, PHI = np.meshgrid(theta, phi)
52
53# Convert to Cartesian for plotting
54X = rho * np.sin(PHI) * np.cos(THETA)
55Y = rho * np.sin(PHI) * np.sin(THETA)
56Z = rho * np.cos(PHI)
57
58fig = plt.figure(figsize=(10, 8))
59ax = fig.add_subplot(111, projection='3d')
60ax.plot_surface(X, Y, Z, alpha=0.6, cmap='viridis')
61ax.set_xlabel('X')
62ax.set_ylabel('Y')
63ax.set_zlabel('Z')
64ax.set_title('Sphere in Spherical Coordinates: rho = 2')
65plt.show()

Integration Example: Comparing Coordinate Systems

1import numpy as np
2from scipy import integrate
3
4def cylinder_volume_cartesian(R, H):
5    """
6    Compute cylinder volume using Cartesian coordinates.
7    This requires complex bounds for x, y integration.
8    """
9    def integrand(z, y, x):
10        return 1
11
12    def y_lower(x): return -np.sqrt(R**2 - x**2)
13    def y_upper(x): return np.sqrt(R**2 - x**2)
14
15    result, _ = integrate.tplquad(
16        integrand,
17        -R, R,           # x bounds
18        y_lower, y_upper, # y bounds (depend on x)
19        0, H             # z bounds
20    )
21    return result
22
23def cylinder_volume_cylindrical(R, H):
24    """
25    Compute cylinder volume using cylindrical coordinates.
26    Much simpler bounds!
27
28    V = ∫∫∫ r dr dθ dz
29    """
30    def integrand(z, theta, r):
31        return r  # The Jacobian factor!
32
33    result, _ = integrate.tplquad(
34        integrand,
35        0, R,            # r bounds: 0 to R
36        0, 2*np.pi,      # theta bounds: 0 to 2π
37        0, H             # z bounds: 0 to H
38    )
39    return result
40
41def sphere_volume_spherical(R):
42    """
43    Compute sphere volume using spherical coordinates.
44
45    V = ∫∫∫ ρ² sin(φ) dρ dθ dφ
46    """
47    def integrand(phi, theta, rho):
48        return rho**2 * np.sin(phi)  # The Jacobian!
49
50    result, _ = integrate.tplquad(
51        integrand,
52        0, R,            # rho bounds: 0 to R
53        0, 2*np.pi,      # theta bounds: 0 to 2π
54        0, np.pi         # phi bounds: 0 to π
55    )
56    return result
57
58# Test the formulas
59R, H = 2, 3
60print(f"Cylinder (R={R}, H={H}):")
61print(f"  Cartesian:   {cylinder_volume_cartesian(R, H):.4f}")
62print(f"  Cylindrical: {cylinder_volume_cylindrical(R, H):.4f}")
63print(f"  Exact (πR²H): {np.pi * R**2 * H:.4f}")
64
65print(f"\nSphere (R={R}):")
66print(f"  Spherical:   {sphere_volume_spherical(R):.4f}")
67print(f"  Exact (4/3πR³): {4/3 * np.pi * R**3:.4f}")

Common Pitfalls

Test Your Understanding

Summary

Cylindrical and spherical coordinates are powerful tools for problems with rotational or radial symmetry. They transform complex boundaries into simple constants and make otherwise intractable integrals manageable.

Key Formulas

Key Takeaways

1. Cylindrical coordinates $(r, \theta, z)$ extend polar coordinates to 3D and are ideal for cylinders and axial symmetry
2. Spherical coordinates $(\rho, \theta, \phi)$ describe points by distance and two angles, perfect for spheres and radial symmetry
3. The Jacobian accounts for how coordinate transformations stretch or compress space — never forget it in integrals!
4. Choose coordinates that match your problem's symmetry: a sphere is $\rho = R$ in spherical but $x^2 + y^2 + z^2 = R^2$ in Cartesian
5. These concepts extend to machine learning through hyperspherical geometry, normalizing flows, and 3D computer vision

Coming Next: In the next chapter, we'll explore Vector-Valued Functions, learning how to describe curves and motion in three-dimensional space using vectors that change with a parameter.