06-Matrix-Plots · Python & Machine Learning

Matrix Plots

NOTE: Make sure to watch the video lecture, not all datasets are well suited for a heatmap or clustermap.

Imports

The Data

World Population Prospects publishes United Nations population estimates for all world countries and every year from 1950 to 2020, as well as projections for different scenarios (low, middle and high variants) from 2020 to 2100. The figures presented here correspond to middle variant projections for the given year.

https://www.ined.fr/en/everything_about_population/data/all-countries/?lst_continent=900&lst_pays=926

Source : Estimates for the current year based on data from the World Population Prospects. United Nations.

python

# 2020 Projections
df = pd.read_csv('country_table.csv')

python

df

Out · table

Countries
Birth rate
Mortality rate
Life expectancy
Infant mortality rate

50 more lines

Growth rate
0
AFRICA
32.577
7.837
63.472
44.215
24.40
1
ASIA
15.796
7.030
73.787
23.185
8.44
2
EUROPE
10.118
11.163
78.740
3.750
0.38
3
LATIN AMERICA AND THE CARIBBEAN
15.886
6.444
75.649
14.570
8.89
4
NORTHERN AMERICA
11.780
8.833
79.269
5.563
6.11
5
OCEANIA
16.235
6.788
78.880
16.939
12.79
6
WORLD
17.963
7.601
72.766
27.492
10.36

Heatmap

python

df = df.set_index('Countries')

python

df

Out · table

Birth rate
Mortality rate
Life expectancy
Infant mortality rate
Growth rate

43 more lines

Countries
AFRICA
32.577
7.837
63.472
44.215
24.40
ASIA
15.796
7.030
73.787
23.185
8.44
EUROPE
10.118
11.163
78.740
3.750
0.38
LATIN AMERICA AND THE CARIBBEAN
15.886
6.444
75.649
14.570
8.89
NORTHERN AMERICA
11.780
8.833
79.269
5.563
6.11
OCEANIA
16.235
6.788
78.880
16.939
12.79
WORLD
17.963
7.601
72.766
27.492
10.36

python

# Clearly shows life expectancy in different units
sns.heatmap(df)

Out · result

<AxesSubplot:ylabel='Countries'>

python

rates = df.drop('Life expectancy',axis=1)

python

sns.heatmap(rates)

Out · result

<AxesSubplot:ylabel='Countries'>

python

sns.heatmap(rates,linewidth=0.5)

Out · result

<AxesSubplot:ylabel='Countries'>

python

sns.heatmap(rates,linewidth=0.5,annot=True)

Out · result

<AxesSubplot:ylabel='Countries'>

python

# Note how its not palette here
sns.heatmap(rates,linewidth=0.5,annot=True,cmap='viridis')

Out · result

<AxesSubplot:ylabel='Countries'>

python

# Set colorbar based on value from dataset
sns.heatmap(rates,linewidth=0.5,annot=True,cmap='viridis',center=40)

Out · result

<AxesSubplot:ylabel='Countries'>

python

# Set colorbar based on value from dataset
sns.heatmap(rates,linewidth=0.5,annot=True,cmap='viridis',center=1)

Out · result

<AxesSubplot:ylabel='Countries'>

Clustermap

Plot a matrix dataset as a hierarchically-clustered heatmap.

python

sns.clustermap(rates)

Out · result

<seaborn.matrix.ClusterGrid at 0x158e27976c8>

python

sns.clustermap(rates,col_cluster=False)

Out · result

<seaborn.matrix.ClusterGrid at 0x158e235c9c8>

python

sns.clustermap(rates,col_cluster=False,figsize=(12,8),cbar_pos=(-0.1, .2, .03, .4))

Out · result

<seaborn.matrix.ClusterGrid at 0x158e2ffc848>

python

rates.index.set_names('',inplace=True)

python

rates

Out · table

Birth rate
Mortality rate
Infant mortality rate
Growth rate
AFRICA

34 more lines

32.577
7.837
44.215
24.40
ASIA
15.796
7.030
23.185
8.44
EUROPE
10.118
11.163
3.750
0.38
LATIN AMERICA AND THE CARIBBEAN
15.886
6.444
14.570
8.89
NORTHERN AMERICA
11.780
8.833
5.563
6.11
OCEANIA
16.235
6.788
16.939
12.79
WORLD
17.963
7.601
27.492
10.36

python

# Recall you can always edit the DF before seaborn
sns.clustermap(rates,col_cluster=False,figsize=(12,8),cbar_pos=(-0.1, .2, .03, .4))

Out · result

<seaborn.matrix.ClusterGrid at 0x158e354b508>