22-matplotlib
November 17, 2020
1 Matplotlib
In this lecture we will talk about how to produce scientific graphs using the python library mat-
plotlib. Matplotlib provides a variety of functions that will allow you to quickly and easily pro-
duce a variety of useful, pretty graphs.
Matplotlib is not included by default withpPython. It is a separate python library that is down-
loaded and installed alongside python. If you installed python using anaconda, you probably
already have it.
In order to use matplotlib you need to import it like this
import [Link] as plt
The name after as is just an abbreviation for the library. That means that, whenever we want to
use a function from matplotlib, we will prefix it with plt.
DISCLAIMER: The graphs we will plot here are highly customizable, and we are far from ex-
hausting every configuration option. We encourage you to check matplotlib’s documentation and
the gallery of examples to find out more.
1.1 Prelude
The code below is similar for all that comes afterwards, so we will keep it once here to avoid
having to copy/paste it everywhere.
[1]: import [Link] as plt
import math
# Generates a range of floating point numbers
def rangeFloat(low, hi, step):
res = []
i = low
while i < hi:
[Link](i)
i += step
return res
1
1.2 Plot
The plot graph simply plots whatever coordinates we pass to it.
If we pass coordinates separately, it plots the points:
[2]: import [Link] as plt
def drawPoints():
# Setting the plot's title
[Link]("My first graph")
# Setting the label for the x and y axes
[Link]("x axis")
[Link]("y axis")
# Plotting each point separately
# 'b' stands for blue, and 'o' stands for circle
[Link](1,1,'bo')
[Link](2,4,'bo')
[Link](3,9,'bo')
[Link](4,16,'bo')
[Link](5,25,'bo')
# Renders the graph on the screen
[Link]()
drawPoints()
2
Alternatively, we can pass the x coordinates as a list (first parameter) and the y coordinates as a
list (second parameter). In this case, mathplotlib will connect the dots. Observe that the lists must
be in the correct order. That means that (x,y) coordinates of the same point need to be at the same
position in both lists.
[3]: def drawLine():
# First parameter: x coordinates
# Second parameter: y coordinates
# 'r' stands for red
[Link]([1,2,3,4,5],[1,4,9,16,25], 'r')
[Link]()
drawLine()
3
If we make the coordinates closer, the line will look smoother. The function below plots a graph
of the sine function, using x coordinates at every 0.1.
[4]: def drawSineGraph():
# Generate a list containing all the values from 0 to 10
# with a step of 0.1. We'll use these as our x-values.
x = rangeFloat(0,10,0.1)
# For each x value, generate the appropriate y-value.
sin_x = []
for i in x:
sin_x += [[Link](i)]
# Plot the x,y values with a blue line
[Link](x, sin_x, 'b')
[Link]("This is the title!")
[Link]("X-Axis Label")
[Link]("Y-Axis Label")
# Show the plot
[Link]()
drawSineGraph()
4
We can plot more than one graph in one.
[5]: def drawSineCosineGraph():
# Generate a list containing all the values from 0 to 10
# with a step of 0.1. We'll use these as our x-values.
x = rangeFloat(0,10,0.1)
# For each x value, generate the appropriate y-value.
sin_x = []
cos_x = []
for i in x:
sin_x += [[Link](i)]
cos_x += [[Link](i)]
# Plot sine with a blue line and cosine with a red line
[Link](x, sin_x, 'b')
[Link](x, cos_x, 'r')
[Link]("This is the title!")
[Link]("X-Axis Label")
[Link]("Y-Axis Label")
# Show the plot
[Link]()
5
drawSineCosineGraph()
1.3 Pie charts
The pie function from matplotlib plots pie charts. It expects a list of values as a parameter, and
it will create one pie slice for each one of those values. The whole pie corresponds to the sum of
values.
Optionally, we can pass a named parameter to this function to determine the label of each slice.
The name of this parameter is labels, and the label must be at the same position in the list as its
slice.
Another useful optional paramter for the pie function is autopct="%.1f%%" which writes the per-
centage of each slice on the pie.
[6]: def drawPie():
[Link](figsize=(7,7)) # Make the chart a perfect square
[Link]("A Pie Chart")
[Link]([20,30,50], labels=["20 label", "30 label", "50 label"], autopct="%.
,→1f%%")
[Link]()
drawPie()
6
1.4 Histograms
The hist function plots histograms. A histogram takes a list of values, organizes them into bins,
and then graphs how many items are in each bin. Note that the order of this list of values does
not matter.
Let’s start with a basic example that takes a small list of numbers and generates a histogram using
the bins [0,2) (2 is not included in the bin), [2,4) (4 is not included in the bin), and [4,6]. Observe
that the bin list specifies where each bin begins, and where the last bin ends.
[7]: def drawHistogram():
L = [1, 2, 3, 4, 4, 4, 2, 2]
[Link]("This is a Histogram")
[Link](L, [0,2,4,6])
7
[Link]()
drawHistogram()
As you can see from the output, there is one number in the range [0,2), four numbers in the range
[2,4), and three numbers in the range [4,6].
Instead of including a list indicating the start and end point of the bins, you can also just specify
the total number of bins you want and matplotlib will automatically generate the bin ranges.
Matplotlib generates the bins by taking the lowest and highest values in the list, and splitting this
interval into the number of bins requested.
[8]: def drawHistogram():
L = [1, 2, 3, 4, 4, 4, 2, 2]
[Link]("This is a Histogram")
[Link](L, 3) # We want three bins.
[Link]()
drawHistogram()
8
There are a number of simple arguments you can pass to hist in order to improve the appearance
of the histogram. From the example below, can you figure out what rwidth and color do? Change
the values and experiment to see what happens to the graph.
[9]: def drawHistogram():
L = [1, 2, 3, 4, 4, 4, 2, 2]
[Link]("This is a Histogram")
[Link](L, 3, rwidth=0.8, color="g")
[Link]()
drawHistogram()
9
1.5 Bar charts
The function bar is used to plot bar charts. A bar chart looks a lot like a histogram, but the differ-
ence is that the “bins” on the x-axis may be completely unrelated categories, while in histograms
these are continuous values.
The bar function takes as paramters a list of the positions of the bars on the x-axis, and a list of the
heights of each bar. Optionally you can define the named parameter labels as the list of labels for
the bars to be placed on the x-axis.
[10]: def drawBarChart():
[Link]("A Bar Chart")
# Position of bars, height of bars
[Link]([0,2,4,6,8], [100,88,75,98,64], tick_label=["label1", "label2",␣
,→"label3", "label4", "label5"])
[Link]()
drawBarChart()
10
1.6 Subplots
Sometimes you want to produce multiple different plots and display all of them at the same time,
but not on the same set of axes. We can accomplish this with subplots.
The subplot function can be used to allow multiple plots to be displayed in the same figure.
It takes as arguments three numbers: the number of rows, the number of columns, and which
subplot you want to activate. For example, [Link](121) says to arrange the subplots in a
grid with 1 row and two columns, and then activate the first subplot. (In the case, the left one.)
[Link](122) say to arrange the subplots the same way, but activate the 2nd subplot.
Consider the following example that graphs both sine and cosine in the same figure, but on dif-
ferent subplots.
[11]: def drawSimpleSubplots():
# Generate a list containing all the values from 0 to 10
# with a step of 0.1. We'll use these as our x-values.
x = rangeFloat(0,10,0.1)
# For each x value, generate the appropriate y-value.
sin_x = []
cos_x = []
for i in x:
sin_x += [[Link](i)]
cos_x += [[Link](i)]
11
# Set the size of the plot
[Link](figsize=(14,4))
# Configure the subplot. The layout as a whole is 1 rows with 2 columns,
# and we are currently plotting the first location.
[Link](121)
# Make the first plot
[Link]("This is the left title!")
[Link]("X-Axis Label (left)")
[Link]("Y-Axis Label (left)")
[Link](x,sin_x,'b')
# Switch the subplot to the 2nd subplot
[Link](122)
# Make the second plot
[Link]("This is the right title!")
[Link]("X-Axis Label (right)")
[Link]("Y-Axis Label (right)")
[Link](x,cos_x,'r')
# This help avoid overlap between the two plots
plt.tight_layout()
# Show the final, combined plot
[Link]()
drawSimpleSubplots()
1.7 Exercise
Use the file [Link] containing information
about animals to build a pie chart showing the division of animals per class. The class is indi-
12
cated in the type column by a number ranging from 1 to 7. The mapping of numbers to classes is:
1. Mammals 2. Birds 3. Reptiles 4. Fish 5. Amphibians 6. Insects 7. Others
[12]: def classPieChart():
file = open("18-files/[Link]")
traits = ['hair',
'feathers',
'eggs',
'milk',
'airborne',
'aquatic',
'predator',
'toothed',
'backbone',
'breathes',
'venomous',
'fins',
'legs', # Numeric {0,2,4,5,6,8}
'tail',
'domestic',
'catsize',
'type'] # Numeric [1..7]
typeNames = ["Mammals", "Birds", "Reptiles", "Fish", "Amphibians",␣
,→"Insects", "Others"]
# Reads the csv file as a dictionary of dictionaries
animals = {}
for line in file:
if not [Link]('#'):
vals = [Link]().split(",")
animal = vals[0]
# Builds the internal dictionary for each animal
d = {}
for i in range(len(vals[1:])):
v = vals[i+1]
# Number of legs is numeric
if i == 12:
v = int(v)
# Save the type with the proper name instead of a code
elif i == 16:
typeIdx = int(v) - 1
v = typeNames[typeIdx]
# All else is boolean
else:
13
v = bool(int(v))
d[traits[i]] = v
animals[animal] = d
# Collects the type column as a list
types = []
for animal in animals:
types += [animals[animal]['type']]
# Counts how many animals of each type
mammals = [Link]('Mammals')
birds = [Link]('Birds')
reptiles = [Link]('Reptiles')
fish = [Link]('Fish')
amphibians = [Link]('Amphibians')
insects = [Link]('Insects')
others = [Link]('Others')
[Link]({'[Link]': 16}) # Makes font bigger
[Link](figsize=(7,7))
[Link]("Representation of classes")
[Link]([mammals, birds, reptiles, fish, amphibians, insects, others],␣
,→labels=typeNames, autopct="%.1f%%")
[Link]()
classPieChart()
14
15
