Week 7 - Other libraries and cool things¶

[1]:

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

seaborn ¶

seaborn is a Python data visualisation library for making statistical graphics. It is built on top of matplotlib and integrates very closely with pandas.

Exploratory visualisations are often much easier with seaborn. For example, with only a few lines of code, we can visualise 5 columns from the titanic dataset.

[2]:

sns.set_theme(context='notebook', style='darkgrid')

df = pd.read_csv('../data/titanic.csv')

ax = sns.relplot(
    data=df,
    x='Age', y='Fare', col='Embarked',
    hue='Sex',
    style='Survived',
    markers={0: 'X', 1: 'o'},
);

Matplotlib figure anatomy¶

A matplotlib figure is a collection of Artist objects stored together in a logical parent-child hierarchy. Here’s a neat way to visualise it.

[3]:

from matplotlib.artist import Artist

# Make a basic example figure
fig, ax = plt.subplots(figsize=(6, 6))
ax.plot(range(100), range(100), label='A diaganol line')
ax.set(
    xlabel='The x-axis',
    ylabel='The y-axis',
    title='Example figure'
)
ax.legend()
ax.annotate(
    text='This is the halfway point',
    xy=(50, 50),
    xytext=(20, 80),
    arrowprops={'width':1, 'facecolor':'k', 'edgecolor':'k'}
)

# A function to plot all of the Artists
def recursive_get_children(artist, depth=0):
    if isinstance(artist, Artist):
        print('  ' * depth + str(artist))
        for child in artist.get_children():
            recursive_get_children(child, depth + 2)

# Call the function on our figure
recursive_get_children(fig)

Figure(600x600)
    Rectangle(xy=(0, 0), width=1, height=1, angle=0)
    AxesSubplot(0.125,0.11;0.775x0.77)
        Line2D(A diaganol line)
        Annotation(50, 50, 'This is the halfway point')
        Spine
        Spine
        Spine
        Spine
        XAxis(75.0,65.99999999999999)
            Text(0.5, 0, 'The x-axis')
            Text(1, 0, '')
            <matplotlib.axis.XTick object at 0x7fdb7b0a8460>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb7b0a8430>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb7b0d0790>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb7b0db8b0>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb68150040>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb68150790>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb68156040>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
            <matplotlib.axis.XTick object at 0x7fdb68150f40>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(0, 1, '')
        YAxis(75.0,65.99999999999999)
            Text(0, 0.5, 'The y-axis')
            Text(0, 0.5, '')
            <matplotlib.axis.YTick object at 0x7fdb7b0a8f10>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb7b0a8ca0>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb7b0db9a0>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb681568b0>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb6815c040>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb6815c790>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb68165040>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
            <matplotlib.axis.YTick object at 0x7fdb68165670>
                Line2D()
                Line2D()
                Line2D()
                Text(0, 0, '')
                Text(1, 0, '')
        Text(0.5, 1.0, 'Example figure')
        Text(0.0, 1.0, '')
        Text(1.0, 1.0, '')
        Legend
            <matplotlib.offsetbox.VPacker object at 0x7fdb7b0d0d00>
                <matplotlib.offsetbox.TextArea object at 0x7fdb7b0d0ee0>
                    Text(0, 0, '')
                <matplotlib.offsetbox.HPacker object at 0x7fdb7b0d0d90>
                    <matplotlib.offsetbox.VPacker object at 0x7fdb7b0d0d30>
                        <matplotlib.offsetbox.HPacker object at 0x7fdb7b0d0d60>
                            <matplotlib.offsetbox.DrawingArea object at 0x7fdb7b0d0850>
                                Line2D(A diaganol line)
                            <matplotlib.offsetbox.TextArea object at 0x7fdb7b0d0820>
                                Text(0, 0, 'A diaganol line')
            FancyBboxPatch((0, 0), width=1, height=1)
        Rectangle(xy=(0, 0), width=1, height=1, angle=0)

Now, to demonstrate the power of matplotlib, let’s traverse this hierarchy in true object-oriented fashion and make some changes to a single element.

[4]:

fig.axes[0].get_xticklabels()[4].set(
    color='r',
    style='italic',
    weight='bold',
    size=42,
    family='Comic Sans MS'
)

fig

[4]:

This may seem like a silly exercise, but it reveals much about matplotlib. What else about the plot can you change?

Animations with matplotlib¶

With matplotlib, it is also possible to make animated plots. Here’s one that shows the number of cycling accidents over time. Note you may need to install some additional libraries for this to work in a Jupyter notebook.

[5]:

df

[5]:

	PassengerID	Survived	Pclass	Name	Sex	Age	SibSp	Parch	Ticket	Fare	Cabin	Embarked
0	1	0	3	Braund, Mr. Owen Harris	male	22.0	1	0	A/5 21171	7.2500	NaN	S
1	2	1	1	Cumings, Mrs. John Bradley (Florence Briggs Th...	female	38.0	1	0	PC 17599	71.2833	C85	C
2	3	1	3	Heikkinen, Miss. Laina	female	26.0	0	0	STON/O2. 3101282	7.9250	NaN	S
3	4	1	1	Futrelle, Mrs. Jacques Heath (Lily May Peel)	female	35.0	1	0	113803	53.1000	C123	S
4	5	0	3	Allen, Mr. William Henry	male	35.0	0	0	373450	8.0500	NaN	S
...	...	...	...	...	...	...	...	...	...	...	...	...
886	887	0	2	Montvila, Rev. Juozas	male	27.0	0	0	211536	13.0000	NaN	S
887	888	1	1	Graham, Miss. Margaret Edith	female	19.0	0	0	112053	30.0000	B42	S
888	889	0	3	Johnston, Miss. Catherine Helen "Carrie"	female	NaN	1	2	W./C. 6607	23.4500	NaN	S
889	890	1	1	Behr, Mr. Karl Howell	male	26.0	0	0	111369	30.0000	C148	C
890	891	0	3	Dooley, Mr. Patrick	male	32.0	0	0	370376	7.7500	NaN	Q

891 rows × 12 columns

[6]:

from matplotlib.animation import FuncAnimation

df = (
    pd.read_csv('../data/gb_cycling_accidents.csv')
    .assign(index=lambda df_: pd.DatetimeIndex(df_.Date + ' ' + df_.Time))
    .set_index('index')
    .assign(Year=lambda df_: df_.index.year)
    .groupby(['Year', 'Gender'])['Accident_Index']
    .count()
    .unstack()
)

fig, ax = plt.subplots(figsize=(8, 4))

ln_male, = ax.plot([], [], 'ro-')
ln_female, = ax.plot([], [], 'bo-')
ln_other, = ax.plot([], [], 'go-')

def init():
    ax.set_ylim(-1000, df.Male.max()*1.05)
    ax.set_xlim((df.index.min(), df.index.max()))
    ax.set_xlabel('Year')
    ax.set_ylabel('Number of accidents')
    ax.set_title('Cycling accidents in Great Britain (1979-2018)')
    ax.legend([ln_male, ln_female, ln_other], ['Males', 'Females', 'Other'])
    return ln_male, ln_female, ln_other,

def update(frame):
    data = df.iloc[0:frame]
    ln_male.set_data(data.index, data.Male)
    ln_female.set_data(data.index, data.Female)
    ln_other.set_data(data.index, data.Other)


    return ln_male, ln_female, ln_other,

ani = FuncAnimation(fig, update, frames=len(df.index.to_numpy()),
                    init_func=init, blit=True)
plt.close()
ani.save('../images/gb_cycling_animation.gif')

3a04453d34644eef9b24adc2032bb1e6

Geographical plots with `cartopy`¶

Map projections:

There are various libraries for plotting geospatial data in Python. A good example is the `cartopy <https://scitools.org.uk/cartopy/docs/latest/>`__ library. Here, I use cartopy to plot the night-time shading for the current time on a flat map of the earth, along with the location of the University of York, and the 10 most populated cities.

The city data are freely available at the following web page:

https://simplemaps.com/data/world-cities

[7]:

import pandas as pd

# Load the city data
df = (
    pd.read_csv('../data/worldcities.csv')
    .sort_values('population', ascending=False)
    .head(15)
)
df

[7]:

	city	city_ascii	lat	lng	country	iso2	iso3	admin_name	capital	population	id
0	Tokyo	Tokyo	35.6839	139.7744	Japan	JP	JPN	Tōkyō	primary	39105000.0	1392685764
1	Jakarta	Jakarta	-6.2146	106.8451	Indonesia	ID	IDN	Jakarta	primary	35362000.0	1360771077
2	Delhi	Delhi	28.6667	77.2167	India	IN	IND	Delhi	admin	31870000.0	1356872604
3	Manila	Manila	14.6000	120.9833	Philippines	PH	PHL	Manila	primary	23971000.0	1608618140
4	São Paulo	Sao Paulo	-23.5504	-46.6339	Brazil	BR	BRA	São Paulo	admin	22495000.0	1076532519
5	Seoul	Seoul	37.5600	126.9900	South Korea	KR	KOR	Seoul	primary	22394000.0	1410836482
6	Mumbai	Mumbai	19.0758	72.8775	India	IN	IND	Mahārāshtra	admin	22186000.0	1356226629
7	Shanghai	Shanghai	31.1667	121.4667	China	CN	CHN	Shanghai	admin	22118000.0	1156073548
8	Mexico City	Mexico City	19.4333	-99.1333	Mexico	MX	MEX	Ciudad de México	primary	21505000.0	1484247881
9	Guangzhou	Guangzhou	23.1288	113.2590	China	CN	CHN	Guangdong	admin	21489000.0	1156237133
10	Cairo	Cairo	30.0444	31.2358	Egypt	EG	EGY	Al Qāhirah	primary	19787000.0	1818253931
11	Beijing	Beijing	39.9040	116.4075	China	CN	CHN	Beijing	primary	19437000.0	1156228865
12	New York	New York	40.6943	-73.9249	United States	US	USA	New York	NaN	18713220.0	1840034016
13	Kolkāta	Kolkata	22.5727	88.3639	India	IN	IND	West Bengal	admin	18698000.0	1356060520
14	Moscow	Moscow	55.7558	37.6178	Russia	RU	RUS	Moskva	primary	17693000.0	1643318494

[8]:

import datetime
import matplotlib.pyplot as plt
import numpy as np
import cartopy.crs as ccrs
from cartopy.feature.nightshade import Nightshade
%matplotlib widget

# Create a figure with a GeoAxes by specifying
fig = plt.figure(figsize=(12, 6))
ax = fig.add_subplot(1, 1, 1, projection=ccrs.PlateCarree())

# Get current date and time
dt = datetime.datetime.now()

# Location of University of York
location = (-1.0311947681813436, 53.94930227196749)

# Arrow props
arrowprops=dict(
    arrowstyle='fancy',
    shrinkA=5,
    shrinkB=5,
    fc="k", ec="k",
    connectionstyle="arc3,rad=-0.05",
)

# Add title
ax.set_title(f'Night time shading for {dt}')

# Draw a standard flat map of the world
ax.stock_img()

# Add the nightshade feature
ax.add_feature(Nightshade(dt, alpha=0.4))

# Add University of York location and annotate
ax.scatter(*location, c='r', s=5)
ax.annotate(
    text='University of York',
    xy=location,
    xytext=(-65, 20),
    arrowprops=arrowprops,
    fontweight='bold'
)

# Plot the city locations
ax.scatter(df.lng, df.lat, c='k', s=5)

#Annotate with the names of the cities
for idx, row in df.iterrows():
    ax.annotate(
        text=row.city,
        xy=(row.lng+1, row.lat+1),
        fontsize=8

    )

plt.tight_layout()
plt.show()

matplotlib has its own Basemap Toolkit which predates cartopy. Soon I’ll be off to Copenhagen, so I decided to use it to plot the great circle route between airports.

[9]:

from mpl_toolkits.basemap import Basemap
import numpy as np
import matplotlib.pyplot as plt

# create new figure, axes instances.
fig=plt.figure(figsize=(12, 4))
ax=fig.add_axes([0.1,0.1,0.8,0.8])

# setup mercator map projection.
m = Basemap(
    llcrnrlon=-15.,llcrnrlat=45.,urcrnrlon=25.,urcrnrlat=65.,
    rsphere=(6378137.00,6356752.3142),
    resolution='l',projection='merc',
    lat_0=40.,lon_0=-20.,lat_ts=20.
)

# lat/lon for manchester and copenhagen
cop_lat, cop_lon = 55.62798787190983, 12.643942953245418
man_lat, man_lon = 53.35544507391249, -2.277185420260674

# draw great circle route between manchster and copenhagen
m.drawgreatcircle(cop_lon,cop_lat,man_lon,man_lat,linewidth=2,color='b')
m.drawcoastlines()
m.fillcontinents()

# draw parallels
m.drawparallels(np.arange(10,90,20), labels=[1,1,0,1])

# draw meridians
m.drawmeridians(np.arange(-180,180,30),labels=[1,1,0,1])

ax.set_title('Great Circle from Manchester to Copenhagen')
plt.show()

Week 7 - Other libraries and cool things¶

seaborn¶

Matplotlib figure anatomy¶

Animations with matplotlib¶

Geographical plots with cartopy¶

seaborn ¶

Geographical plots with `cartopy`¶