LineCollection and PolyCollection for speed in Matplotlib - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of LineCollection plot segments count

What is the number of line segments drawn by this code using LineCollection?

Matplotlib

import matplotlib.pyplot as plt
from matplotlib.collections import LineCollection

segments = [ [(0, 0), (1, 1)], [(1, 1), (2, 0)], [(2, 0), (3, 1)]]
line_collection = LineCollection(segments)
fig, ax = plt.subplots()
ax.add_collection(line_collection)
ax.autoscale()
plt.close(fig)

num_segments = len(line_collection.get_segments())
print(num_segments)

A 1

B 2

C 3

D 0

💡 Hint

Count how many pairs of points are in the segments list.

LineCollection takes a list of line segments. Each segment is a pair of points. Here, there are 3 segments, so the output is 3.

❓ data_output

intermediate

Number of polygons in PolyCollection

How many polygons does this PolyCollection contain?

Matplotlib

from matplotlib.collections import PolyCollection

polygons = [ [(0, 0), (1, 0), (0.5, 1)], [(1, 1), (2, 1), (1.5, 2)]]
poly_collection = PolyCollection(polygons)
num_polygons = len(poly_collection.get_paths())
print(num_polygons)

A 1

B 2

C 0

D 3

💡 Hint

Each polygon is a list of points; count how many polygons are given.

PolyCollection holds multiple polygons. Here, two polygons are passed, so the output is 2.

❓ visualization

advanced

Visual difference between LineCollection and multiple plot calls

Which option shows the fastest way to plot 1000 line segments using matplotlib?

Matplotlib

import matplotlib.pyplot as plt
from matplotlib.collections import LineCollection
import numpy as np

segments = [ [(i, 0), (i, 1)] for i in range(1000)]

# Option A: Using LineCollection
fig, ax = plt.subplots()
lc = LineCollection(segments)
ax.add_collection(lc)
ax.autoscale()
plt.close(fig)

# Option B: Using 1000 plot calls
fig2, ax2 = plt.subplots()
for seg in segments:
    ax2.plot([seg[0][0], seg[1][0]], [seg[0][1], seg[1][1]])
plt.close(fig2)

A Neither method works for large numbers of segments.

B Using 1000 plot calls is faster because each line is drawn separately.

C Both methods have the same speed because matplotlib optimizes internally.

D Using LineCollection is faster because it batches all segments in one object.

💡 Hint

Think about how matplotlib handles many lines internally.

LineCollection batches all line segments into one object, reducing overhead and speeding up rendering compared to many individual plot calls.

🔧 Debug

advanced

Error when passing wrong data to PolyCollection

What error does this code raise when creating a PolyCollection with incorrect polygon data?

Matplotlib

from matplotlib.collections import PolyCollection

polygons = [ [(0, 0), (1, 0)], [(1, 1), (2, 1), (1.5, 2)]]
poly_collection = PolyCollection(polygons)

A ValueError: Polygons must have at least 3 vertices

B TypeError: Expected a list of tuples

C IndexError: list index out of range

D No error, PolyCollection accepts any polygon

💡 Hint

A polygon needs at least 3 points to be valid.

PolyCollection requires each polygon to have at least 3 vertices. The first polygon has only 2 points, causing a ValueError.

🚀 Application

expert

Efficiently plotting thousands of polygons with PolyCollection

You want to plot 5000 polygons efficiently. Which approach is best?

Matplotlib

import matplotlib.pyplot as plt
from matplotlib.collections import PolyCollection
import numpy as np

polygons = [np.random.rand(4, 2) for _ in range(5000)]

# Option A: Create one PolyCollection with all polygons
# Option B: Plot each polygon with plt.fill()
# Option C: Use multiple PolyCollections each with 100 polygons
# Option D: Use scatter plot instead

A Create one PolyCollection with all polygons to minimize draw calls.

B Plot each polygon individually with plt.fill() for better control.

C Use multiple PolyCollections with smaller batches to avoid memory issues.

D Use scatter plot to represent polygons as points.

💡 Hint

Batching many polygons reduces overhead and speeds up rendering.

One PolyCollection with all polygons reduces the number of draw calls and is the most efficient way to plot many polygons.