Bar Charts

This notebook introduces the methods of creating bar charts from our data, including some of the key concepts of our plotting package, matplotlib. Bar charts are useful for making comparisons between close data because we are very good at evaluating the size of rectangles.

Topics covered:

• bar charts

• styling charts (titles, labels, size, color, etc)

import matplotlib as plt
from nycschools import schools

df = schools.load_school_demographics()

Simple Bar Chart

We can quickly and easily create charts in pandas by calling the bar.plot() function on a DataFrame. The first step is to create a table where the data is ready to present. In this example, we’ll create a bar chart from the school enrollment data in df (the school demographics).

We’ll get started by borrowing the code from our groupby and agg examples.

# get a few columns for just the 2020-21 school year
boros = df[["boro", "dbn", "total_enrollment", "ay"]]
boros = boros[boros.ay == 2020]

# group the data by borough
table = boros.groupby(by="boro").agg({
    "dbn":"count", 
    "total_enrollment":["sum", "mean", "min", "max"]}).reset_index()

# rename the columns with "pretty" names
cols = ["Borough", "Num. Schools", "Total Students", "Avg School Size", "Smallest School", "Largest School"]
table.columns = cols
table

	Borough	Num. Schools	Total Students	Avg School Size	Smallest School	Largest School
0	Bronx	496	238074	479.987903	7	2936
1	Brooklyn	640	346802	541.878125	46	5921
2	Manhattan	408	194144	475.843137	14	3342
3	Queens	421	305660	726.033254	42	4424
4	Staten Island	83	63824	768.963855	144	3726

Drawing the plot

Making a chart from our table is as simple as calling plot.bar(). In the example below we specify that we want enrollment (school size) as the vertical y-axis and Borough as our x-axis.

ax = table.plot.bar(x="Borough", y=["Total Students"])

# doe the same thing with average school size
ax = table.plot.bar(x="Borough", y=["Avg School Size"])

Labels, figure size, and more

We can make a pretty decent bar chart for a single line of code, but with a few more arguments to our function, we can make it even better. To do this, we’re going to use a few more features from matplotlib and related libraries.

In this example we will specify:

• sorting the columns by size

• setting the images size

• adding titles and labels

• changing colors

You may have noticed that we were assigning the return value from plot() into a new variable called ax. This servers the simple purpose of showing only the chart (and not the return value), but it also allows us to keep adding items to our chart.

# before drawing anything, change the "style"
plt.style.use('dark_background')

# show all the styles
# change dark_background to any style from this list to try it out
plt.style.available

['Solarize_Light2',
 '_classic_test_patch',
 '_mpl-gallery',
 '_mpl-gallery-nogrid',
 'bmh',
 'classic',
 'dark_background',
 'fast',
 'fivethirtyeight',
 'ggplot',
 'grayscale',
 'seaborn-v0_8',
 'seaborn-v0_8-bright',
 'seaborn-v0_8-colorblind',
 'seaborn-v0_8-dark',
 'seaborn-v0_8-dark-palette',
 'seaborn-v0_8-darkgrid',
 'seaborn-v0_8-deep',
 'seaborn-v0_8-muted',
 'seaborn-v0_8-notebook',
 'seaborn-v0_8-paper',
 'seaborn-v0_8-pastel',
 'seaborn-v0_8-poster',
 'seaborn-v0_8-talk',
 'seaborn-v0_8-ticks',
 'seaborn-v0_8-white',
 'seaborn-v0_8-whitegrid',
 'tableau-colorblind10']

# sort the tabgle in descending order by school size -- this will order our bars
table = table.sort_values(by="Avg School Size", ascending=False)

ax = table.plot.bar(x="Borough", y=["Avg School Size"], color="violet")
ax.set_title("Average School Size by Borough (AY 2020-21)")
ax.set_ylabel("Number of Students")
ax.get_legend().remove()