Visualisation¶

Non-spatial visualisation
Spatial visualisation
Choropleth mapping
Anatomy of a graphic
- Figures
- Axes
- Layers
Exercises

In [1]:

%matplotlib inline

import geopandas
import seaborn
import contextily
import matplotlib.pyplot as plt

db = geopandas.read_file('../data/lux_regions.gpkg')

Non-spatial visualisation¶

Univariate continuous¶

Histograms

In [2]:

db['POPULATION'].plot.hist()

Out[2]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc591022a58>

KDEs

In [3]:

db['light_level'].plot.kde()

Out[3]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc590eed2e8>

Bivariate continuous¶

Scatter plots

In [4]:

db[['POPULATION', 'light_level']]\
  .plot.scatter('light_level', 
                'POPULATION')

Out[4]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc590e5de80>

Hexbin plots

In [5]:

seaborn.jointplot('light_level', 
                  'POPULATION', 
                  db, 
                  kind='hex')

Out[5]:

<seaborn.axisgrid.JointGrid at 0x7fc590e495c0>

2D KDEs

In [6]:

seaborn.jointplot('light_level', 
                  'POPULATION', 
                  db, 
                  kind='kde')

Out[6]:

<seaborn.axisgrid.JointGrid at 0x7fc58ec8d0b8>

Categorical plots¶

Categorical scatter plots

In [7]:

seaborn.catplot(x="DISTRICT",
                y="POPULATION",
                data=db)

Out[7]:

<seaborn.axisgrid.FacetGrid at 0x7fc58eb76908>

Box plots

In [8]:

seaborn.catplot(x="DISTRICT",
                y="POPULATION",
                data=db,
                kind='box')

Out[8]:

<seaborn.axisgrid.FacetGrid at 0x7fc58ed8db38>

Violin plots

In [9]:

seaborn.catplot(x="DISTRICT",
                y="POPULATION",
                data=db,
                kind='violin')

Out[9]:

<seaborn.axisgrid.FacetGrid at 0x7fc58eaed198>

Spatial visualisation¶

In [10]:

db.geometry.head()

Out[10]:

0    POLYGON ((5.999664443592493 49.51861834298563,...
1    POLYGON ((6.347791382484162 49.76015684859829,...
2    POLYGON ((5.966652609366625 49.51287419342548,...
3    POLYGON ((6.125796013396808 49.66347171771235,...
4    POLYGON ((6.08267171189537 49.77458013764257, ...
Name: geometry, dtype: object

In [11]:

db.plot()

Out[11]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc58e9ba898>

In [12]:

ax = db.plot()
contextily.add_basemap(ax, crs=db.crs)

Choropleth mapping¶

In [13]:

db.plot(column="POPULATION")

Out[13]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc58e8fa0b8>

Classification schemes¶

Unique vales (categorical)

In [14]:

db.plot(column="CANTON", 
        categorical=True)

Out[14]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc58cd3b860>

Equal intervals

In [15]:

db.plot(column="POPULATION", 
        scheme='equal_interval',
        legend=True)

Out[15]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc58e570be0>

Quantiles

In [16]:

db.plot(column="POPULATION", 
        scheme='quantiles',
        legend=True)

Out[16]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc583e75b70>

Fisher-Jenks

In [17]:

db.plot(column="POPULATION", 
        scheme='fisher_jenks',
        legend=True)

Out[17]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc583dd52b0>

Number of classes¶

In [18]:

db.plot(column="POPULATION", 
        scheme='quantiles',
        k=3,
        legend=True)

Out[18]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc58317c470>

In [19]:

db.plot(column="POPULATION", 
        scheme='quantiles',
        k=10,
        legend=True)

Out[19]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc583133c18>

Palettes¶

[See matplotlib colormaps]

In [20]:

db.plot(column="POPULATION", 
        scheme='quantiles',
        cmap='plasma')

Out[20]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc583002320>

Categorical

In [21]:

db.plot(column="CANTON",
        cmap='Pastel1')

Out[21]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc58303ca58>

Sequential

In [22]:

db.plot(column="POPULATION", 
        scheme='quantiles',
        cmap='RdPu')

Out[22]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc582fb76a0>

Divergent

In [23]:

z = (db['tree_count'] - db['tree_count'].median()) /\
     db['tree_count'].std()

db.assign(z=z)\
  .plot(column="z",
        scheme="quantiles",
        cmap="bwr",
        legend=True)

Out[23]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc582e40278>

Anatomy of a graphic¶

Figures¶

In [24]:

f = plt.figure()

<Figure size 432x288 with 0 Axes>

Change size

In [25]:

# Size
f = plt.figure(figsize=(12, 12))

<Figure size 864x864 with 0 Axes>

Axes¶

One axis (ax) inside a figure (f)

In [26]:

f, ax = plt.subplots(1)

Two rows, one column of axes (axs) inside a figure (f)

In [27]:

f, axs = plt.subplots(2)

One row, two columns of axes (axs) inside a figure (f)

In [28]:

f, axs = plt.subplots(1, 2)

Embed data on an axis

In [29]:

f, ax = plt.subplots(1)
db.plot(ax=ax)

Out[29]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc581c61e80>

Embed data on two axes

In [30]:

f, axs = plt.subplots(1, 2, figsize=(12, 6))
# First axis
db.plot(column='POPULATION',
        scheme='quantiles',
        ax=axs[0])
# Second axis
db['POPULATION'].plot.kde(ax=axs[1])
# Title
f.suptitle("Population")
# Display
plt.show()

Layers¶

Tweak a layer

In [31]:

# Transparency
db.plot(alpha=0.5)

Out[31]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc581c1b6d8>

In [32]:

# Color
db.plot(color="red")

Out[32]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fc583507ef0>

Exercises¶

Create a scatter plot using seaborn's jointplot (hint: have a look at the seaborn's tutorial for plotting bivariate distributions)
Replicate the Boxplot using the boxen option for a cooler alternative
Create a map of Luxembourg regions with the following characteristics:
- 50% transparency
- A basemap
- Orange polygons
Create a figure with three subplots displaying the following:
- Histogram of the light_level
- Scatter plot of light_level against tree_count
- KDE of the tree_count