How to format the x-axis by hour increments

Question

I'm working with a dataset that includes the bedtimes of participants for weekdays, weekends, and the entire week. The bedtimes are represented as strings in the "hh:mm" format. Here's a sample of the data:

Au cours de la semaine	En weekend	Toute la semaine
22:30:00	24:30:00	23:30:00
21:30:00	23:00:00	22:15:00
23:00:00	24:30:00	23:45:00
23:30:00	27:00:00	25:15:00
24:30:00	25:30:00	25:00:00
21:45:00	21:45:00	21:45:00
23:15:00	24:15:00	23:45:00
22:30:00	24:30:00	23:30:00
23:45:00	24:30:00	24:07:30
20:30:00	21:30:00	21:00:00
22:30:00	24:00:00	23:15:00
missing	missing	missing
21:15:00	21:00:00	21:07:30
22:30:00	24:30:00	23:30:00
23:45:00	26:30:00	25:07:30
21:30:00	23:30:00	22:30:00

I'm struggling to fix the x-axis intervals to display time ticks at a fixed 1-hour interval. Here's the code I used :

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

data = {'Au cours de la semaine': ['22:30:00', '21:30:00', '23:00:00', '23:30:00', '24:30:00', '21:45:00', '23:15:00', '22:30:00', '23:45:00', '20:30:00', '22:30:00', 'missing', '21:15:00', '22:30:00', '23:45:00', '21:30:00'],
        'En weekend': ['24:30:00', '23:00:00', '24:30:00', '27:00:00', '25:30:00', '21:45:00', '24:15:00', '24:30:00', '24:30:00', '21:30:00', '24:00:00', 'missing', '21:00:00', '24:30:00', '26:30:00', '23:30:00'],
        'Toute la semaine': ['23:30:00', '22:15:00', '23:45:00', '25:15:00', '25:00:00', '21:45:00', '23:45:00', '23:30:00', '24:07:30', '21:00:00', '23:15:00', 'missing', '21:07:30', '23:30:00', '25:07:30', '22:30:00']}
df = pd.DataFrame(data).replace('missing', pd.NA).dropna()

# Convert time data to minutes
for column in df.columns:
    df[column] = df[column].apply(lambda x: int(x.split(':')[0])*60 + int(x.split(':')[1]))

# Melt the DataFrame
df_melt = pd.melt(df, var_name='Category', value_name='Time')

# Create a function to format x-axis ticks
def format_func(value, tick_number):
    # Convert minutes to hours and minutes
    hours = int(value // 60)
    minutes = int(value % 60)
    return f"{hours}:{minutes:02d}"

# Create the boxplots
fig, ax = plt.subplots(figsize=(10,6))
sns.boxplot(x="Time", y="Category", data=df_melt, orient='h', ax=ax, color='lightblue')

# Add means as diamond markers
sns.pointplot(x="Time", y="Category", data=df_melt, orient='h', 
              ax=ax, color='darkred', markers='D', join=False, errorbar=None)

# Format the x-axis ticks
ax.xaxis.set_major_formatter(ticker.FuncFormatter(format_func))

# Set the labels
plt.xlabel("Time (HH:MM)")
plt.title("Boxplots of Bedtime")

# Display the plot
plt.show()

Can anyone suggest a way to modify the code to ensure the x-axis ticks have a fixed 1-hour interval between them?

Answer 1

• Instead of pointplot for the mean marker, use showmeans=True and meanprops=dict(marker='D', markeredgecolor='darkred', markerfacecolor='darkred'), which are described in Artist customization in box plots
• See other comments inline.

import math  # from standard library
import matplotlib.ticker as tkr
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd

# preceded by other code from OP

# Create the boxplots with custom mean properties
fig, ax = plt.subplots(figsize=(10,6))
sns.boxplot(x="Time", y="Category", data=df_melt, orient='h', ax=ax, color='lightblue',
            showmeans=True, meanprops=dict(marker='D', markeredgecolor='darkred', markerfacecolor='darkred'))

# determine the min and max time; maybe buffer the end points by ± 15 minues
time_min = df_melt.Time.min()
time_max = df_melt.Time.max() + 15

# calulate where the x limits should be so they're on the hour
xlim_left = (time_min//60)*60
xlim_right = math.ceil(time_max/60)*60

# set the limits
ax.set_xlim(xlim_left, xlim_right)

# set the new xticks at 60 minute intervals
ax.set_xticks(list(range(xlim_left, xlim_right+1, 60)))

# Format the x-axis ticks
ax.xaxis.set_major_formatter(tkr.FuncFormatter(format_func))

# Set the labels
ax.set(xlabel='Time (HH:MM)', title='Boxplots of Bedtime')

# Display the plot
plt.show()