data_visualization_bar.py

import pickle
import matplotlib.pyplot as plt
import pandas as pd
import networkx as nx

# Load the data from the pkl file
with open("pkl_files/classifications_with_names.pkl", "rb") as file:
    data = pickle.load(file)

# Extract class names from the dictionary
classifications = [value["class_name"] for value in data.values()]

# Count occurrences of each class
class_counts = pd.Series(classifications).value_counts()

# Limit to top 30 classes for better visualization
top_n = 50
filtered_class_counts = class_counts.head(top_n)

# Plot the data as a horizontal bar chart
plt.figure(figsize=(12, 8))
filtered_class_counts.plot(
    kind="barh",
    color=plt.cm.viridis_r(filtered_class_counts / max(filtered_class_counts)),
)

# Add titles and labels
plt.title(f"Top {top_n} Image Classification Counts", fontsize=16)
plt.xlabel("Number of Images", fontsize=14)
plt.ylabel("Class Name", fontsize=14)

# Add counts on the bars
for index, value in enumerate(filtered_class_counts):
    plt.text(
        value, index, str(value), va="center", ha="left", fontsize=12, color="black"
    )

# Improve layout
plt.tight_layout()

# Show the plot
plt.show()