Plotting Categorical Data With Pandas and Matplotlib

Plotting categorical data with pandas and matplotlib

You can simply use value_counts on the series:

df['colour'].value_counts().plot(kind='bar')

How to get a grouped bar plot of categorical data

Imports and Sample Data

import pandas as pd
import seaborn as sns
import numpy as np  # for test data only

np.random.seed(365)
rows = 100
data = {'year': np.random.choice(range(2016, 2021), size=rows),
        'school': np.random.choice(['a', 'b', 'c', 'd', 'e'], size=rows)}
df = pd.DataFrame(data)

# display(df.head())
   year school
0  2018      a
1  2020      b
2  2017      b
3  2019      b
4  2020      c

With seaborn.countplot

# plot and add annotations
p = sns.countplot(data=df, x='year', hue='school')
p.legend(title='School', bbox_to_anchor=(1, 1), loc='upper left')

for c in p.containers:
    # set the bar label
    p.bar_label(c, fmt='%.0f', label_type='edge')

With pandas.DataFrame.plot

• In order to plot the dataframe directly, use pandas.DataFrame.pivot_table to reshape the dataframe and get the 'size' of each group.

dfp = df.pivot_table(index='year', columns='school', values='school', aggfunc='size')

ax = dfp.plot(kind='bar', rot=0)

ax.legend(title='School', bbox_to_anchor=(1, 1), loc='upper left')

for c in ax.containers:
    # set the bar label
    ax.bar_label(c, fmt='%.0f', label_type='edge')

• The following transformations also work
  
  • pandas.DataFrame.groupby & pandas.DataFrame.pivot
  • pandas.crosstab

# groupby and pivot
ax = df.groupby(['year']).school.value_counts().reset_index(name='counts').pivot(index='year', columns='school', values='counts').plot(kind='bar')

# crosstab
ax = pd.crosstab(df.year, df.school).plot(kind='bar')

Plotting pandas dataframe with boolean categorical time-series data

You may want to visualize the data as heatmap.

Code:

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

df = pd.DataFrame()
df['date'] = pd.date_range(start='2021-02-01', end='2021-03-14', freq='D')
df['specimen1'] = 0
df['specimen2'] = 0
df['specimen3'] = 0

df['specimen1'].loc[(df.date >= '2021-02-01') & (df.date <= '2021-02-03')] = 1
df['specimen3'].loc[(df.date >= '2021-02-02') & (df.date <= '2021-02-05')] = 1
df['specimen2'].loc[(df.date >= '2021-02-10') & (df.date <= '2021-02-20')] = 1

df['date'] = df['date'].dt.date
df = df.set_index('date')

# Visualize the data as heatmap
plt.rcParams['font.size'] = 14
fig, ax = plt.subplots(1, 1, figsize=(9, 16))
sns.heatmap(df, cmap=sns.light_palette('lightblue'), cbar=False, ax=ax)
ax.set_xlabel('Specimen')
ax.set_ylabel('Date')
ax.set_yticks([i for i in range(len(df))], [i for i in df.index.values])

plt.show()

# Save the figure
# fig.savefig('out.png', bbox_inches='tight', facecolor='white')

Figure:

how to plot categorical and continuous data in pandas/matplotlib/seaborn

You could do something like this:

Plotting histogram using seaborn for a dataframe

Personally i prefer seaborn for this kind of plots, because it's easier. But you can use matplotlib too.

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

# sample data
samples = 300
ids = range(samples)
gind = np.random.randint(0, 4, samples)
years = np.random.randint(1990, 2000, samples)

# create sample dataframe
gkeys = {1: 'Drama', 2: 'Comedy', 3: 'Action', 4: 'Adventure', 0: 'Thriller'}
df = pd.DataFrame(zip(ids, gind, years),
                  columns=['ID', 'Genre', 'Year'])
df['Genre'] = df['Genre'].replace(gkeys)

# count the year groups
res = df.groupby(['Year', 'Genre']).count()
res = res.reset_index()

# only the max values
# res_ind = res.groupby(['Year']).idxmax()
# res = res.loc[res_ind['ID'].tolist()]

# viz
sns.set(style="white")
g = sns.catplot(x='Year',
                y= 'ID',
                hue='Genre',
                data=res,
                kind='bar',
                ci=None,
                   )
g.set_axis_labels("Year", "Count")
plt.show()

If this are to many bins in a plot, just split it up.

Matplotlib not respecting Pandas categorical value order

Matplotlib doesn't care about Categorical dtype. You should sort your dataframe first by SIZE:

fig, ax = plt.subplots()
df = df.sort_values('SIZE')
ax.scatter(df.SIZE, df.VALUE)
plt.show()

---

Related Topics