Got it, let's walk through how to build that bar plot with Seaborn, matching the style you had in your R script. We'll use the hue parameter to differentiate between the two days (Lundi and Mardi) and structure everything step by step.

Step 1: Set Up Dependencies & Data #

First, we'll need to import the required libraries and load your data into a pandas DataFrame—Seaborn plays nicely with pandas, so this is the first logical step.

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

# Your raw data converted to a list of lists
data = [
    [0, 0, 270.5, "Lundi"],
    [0, 1, 275.0, "Lundi"],
    [0, 2, 270.0, "Lundi"],
    [0, 3, 273.0, "Lundi"],
    [0, 4, 268.0, "Lundi"],
    [0, 5, 272.0, "Lundi"],
    [0, 6, 269.0, "Lundi"],
    [0, 7, 270.0, "Lundi"],
    [0, 8, 269.0, "Lundi"],
    [0, 9, 274.0, "Lundi"],
    [0, 10, 274.0, "Lundi"],
    [0, 11, 274.0, "Lundi"],
    [0, 12, 273.0, "Lundi"],
    [0, 13, 271.0, "Lundi"],
    [0, 14, 377.0, "Lundi"],
    [0, 15, 2685.0, "Lundi"],
    [0, 16, 2654.0, "Lundi"],
    [0, 17, 2706.0, "Lundi"],
    [0, 18, 2751.0, "Lundi"],
    [0, 19, 2325.0, "Lundi"],
    [0, 20, 274.0, "Lundi"],
    [0, 21, 273.0, "Lundi"],
    [0, 22, 277.0, "Lundi"],
    [0, 23, 276.0, "Lundi"],
    [1, 0, 272.5, "Mardi"],
    [1, 1, 274.0, "Mardi"],
    [1, 2, 266.0, "Mardi"],
    [1, 3, 266.0, "Mardi"],
    [1, 4, 267.0, "Mardi"],
    [1, 5, 274.0, "Mardi"],
    [1, 6, 269.0, "Mardi"],
    [1, 7, 266.0, "Mardi"],
    [1, 8, 270.0, "Mardi"],
    [1, 9, 1267.0, "Mardi"],
    [1, 10, 2618.0, "Mardi"],
    [1, 11, 2610.0, "Mardi"],
    [1, 12, 2629.0, "Mardi"],
    [1, 13, 2248.0, "Mardi"],
    [1, 14, 1897.0, "Mardi"]
]

# Convert to a structured DataFrame
df = pd.DataFrame(data, columns=["DAY", "HOUR", "VALEUR_POINT", "NAMEDAY"])

Step 2: Build the Seaborn Bar Plot with Hue #

Now we'll use sns.barplot() with the hue parameter to group bars by day name. We'll also tweak the plot to make it readable—note that some VALEUR_POINT values are extremely large, which will skew the y-axis (we'll cover fixing that below).

# Set a clean style that aligns with common R plot aesthetics
sns.set_style("whitegrid")

# Create a larger figure for better readability
plt.figure(figsize=(14, 8))
ax = sns.barplot(
    x="HOUR", 
    y="VALEUR_POINT", 
    hue="NAMEDAY", 
    data=df,
    palette="Set2"  # Swap this for any Seaborn palette to match your R plot's colors
)

# Add descriptive labels and a title
ax.set_title("VALEUR_POINT by Hour and Day", fontsize=16)
ax.set_xlabel("Hour of Day", fontsize=12)
ax.set_ylabel("VALEUR_POINT", fontsize=12)

# Move the legend to a non-intrusive position
ax.legend(title="Day", loc="upper left", bbox_to_anchor=(1, 1))

# Adjust layout to prevent label cutoff
plt.tight_layout()

# Display the plot
plt.show()

Quick Note on Your Data #

You'll notice the y-axis is dominated by those extremely high VALEUR_POINT values (like 2685 for Lundi, 2618 for Mardi). If these are data entry errors, you can filter them out before plotting with a quick line of code:

# Filter out values above a reasonable threshold (e.g., 500)
filtered_df = df[df["VALEUR_POINT"] < 500]

Just use filtered_df in the sns.barplot() call to get a clearer view of the typical hourly values.

内容的提问来源于stack exchange，提问作者aymeric d