import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

%config InlineBackend.figure_format ='retina'
%load_ext nb_black

Basics

pyplot vs object-oriented interface

Matplotlib has two interfaces: the object-oriented interface renders Axes instances on Figure instances, while the less flexible state-based MATLAB style-interface keeps track of the current figure and axes and other objects and directs plotting functions accordingly (more here).

Object-oriented plot

df = sns.load_dataset("diamonds")

fig, ax = plt.subplots()
ax.scatter(x="carat", y="price", data=df)
ax.set(xlabel="Carat", ylabel="Price")

[Text(0.5, 0, 'Carat'), Text(0, 0.5, 'Price')]

pyplot version

plt.scatter(x="carat", y="price", data=df)
plt.xlabel("Carat")
plt.ylabel("Price")

Text(0, 0.5, 'Price')

Plot lifecycle

Based on this great blog post by Chris Moffitt and the matplotlib tutorial that's based on the same post.

Reading in raw data of customer sales transactions and keeping sales volume and number of purchases for top 10 customers by sales.

fp = (
    "https://github.com/chris1610/pbpython/blob/master/data/"
    "sample-salesv3.xlsx?raw=true"
)
df_raw = pd.read_excel(fp)
print(df_raw.shape)
df_raw.head(2)

(1500, 7)

	account number	name	sku	quantity	unit price	ext price	date
0	740150	Barton LLC	B1-20000	39	86.69	3380.91	2014-01-01 07:21:51
1	714466	Trantow-Barrows	S2-77896	-1	63.16	-63.16	2014-01-01 10:00:47

df = (
    df_raw.groupby("name")
    .agg(sales=("ext price", "sum"), purchases=("quantity", "count"))
    .sort_values("sales")[-10:]
    .reset_index()
)
df

	name	sales	purchases
0	Keeling LLC	100934.30	74
1	Frami, Hills and Schmidt	103569.59	72
2	Koepp Ltd	103660.54	82
3	Will LLC	104437.60	74
4	Barton LLC	109438.50	82
5	Fritsch, Russel and Anderson	112214.71	81
6	Jerde-Hilpert	112591.43	89
7	Trantow-Barrows	123381.38	94
8	White-Trantow	135841.99	86
9	Kulas Inc	137351.96	94

Choosing a style

plt.style.available

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

plt.style.use("seaborn-whitegrid")

Prototyping plot with Pandas

df.plot(kind="barh", x="name", y="sales", legend=None);

Customising plot combining fast Pandas plotting with Matplotlib object-oriented API

def xlim(x):
    """Set xlim with custom padding."""
    return x.max() * np.array([-0.05, 1.3])


fig, ax = plt.subplots()
df.plot(kind="barh", x="name", y="sales", legend=None, ax=ax)
ax.set(
    xlim=xlim(df.sales),
    xlabel="Sales",
    ylabel="Customer",
    title="Top customers 2014",
);

Formatting currency values using custom formatter

def currency(x, pos):
    """Reformat currency amount at position x."""
    return f"{x * 1e-3:1.1f}K"


ax.xaxis.set_major_formatter(currency)

fig

Adding a line for average sales

sales_mean = df.sales.mean()
ax.axvline(sales_mean, linestyle=":", color="green")
lab = f"Mean: {currency(sales_mean, 0)}"
ax.text(
    x=1.05 * sales_mean,
    y=0,
    s=lab,
    color="green",
)

fig

Identify new customers

for customer in [2, 4, 5]:
    ax.text(x=1.05 * sales_mean, y=customer, s="New customer")
fig

Show sales and number of purchases, xkcd-themed (just because...)

with plt.xkcd():
    fig, (ax0, ax1) = plt.subplots(1, 2, figsize=(8, 4), sharey=True)

    df.plot(kind="barh", x="name", y="sales", legend=None, ax=ax0)
    sales_mean = df.sales.mean()
    ax0.axvline(sales_mean, color="green", linestyle=":")
    lab = f"Mean: {currency(sales_mean, 0)}"
    ax0.text(1.05 * sales_mean, 0, lab, color="green")
    for customer in [2, 4, 5]:
        ax0.text(sales_mean, customer, "New customer")
    ax0.xaxis.set_major_formatter(currency)
    ax0.set(xlim=xlim(df.sales), ylabel="Customer",title="Sales")

    df.plot(kind="barh", x="name", y="purchases", legend=None, ax=ax1)
    purch_mean = df.purchases.mean()
    ax1.axvline(purch_mean, color="green", linestyle=":")
    ax1.text(purch_mean, 0, f"Mean: {purch_mean}", color="green")
    ax1.set(title="Purchases", xlim=xlim(df.purchases))

    fig.suptitle(
        "Sales and purchases for top 10 customers in 2022",
        fontsize=18,
        fontweight="bold",
        y=1.05,
    )