app.py

import pandas as pd
import seaborn as sns
import streamlit as st
import matplotlib.pyplot as plt
import plotly.graph_objects as go
from sklearn.cluster import KMeans
df = pd.read_csv("Spotify-2000.csv")


# Main content
st.title('Music Genres Clustering :saxophone:')
st.image('https://www.cnet.com/a/img/resize/034ecacea4dbf0ae76529f37c2d0309b17b557ec/hub/2021/11/10/ab5e2d3b-9a4a-41f0-b2cd-6cee804ce823/genre-charts-covers.png?auto=webp&fit=crop&height=675&width=1200')
st.write('**Data Preview:**')
df = df.drop("Index", axis=1)
st.dataframe(df)

# GENRE CLUSTER
st.write('**Visualizing Top 5 Genres:**')
top_genres = df["Top Genre"].value_counts().nlargest(5).index
def plot_3d_scatter(df):
    fig = go.Figure()
    for i in top_genres:
        fig.add_trace(go.Scatter3d(x=df[df["Top Genre"]==i]['Beats Per Minute (BPM)'],
                                   y=df[df["Top Genre"]==i]['Energy'],
                                   z=df[df["Top Genre"]==i]['Danceability'],
                                   mode='markers', marker=dict(size=6, line=dict(width=1)), name=str(i)))
    fig.update_traces(hovertemplate='BPM: %{x} <br>Energy: %{y} <br>Danceability: %{z}')
    fig.update_layout(autosize=True, scene=dict(xaxis_title='BPM', yaxis_title='Energy', zaxis_title='Danceability'))
    return fig
st.plotly_chart(plot_3d_scatter(df))


# KMEANS CLUSTER
st.write('**Clustering by "Beats Per Minute (BPM)", "Loudness (dB)", "Liveness", "Valence", "Acousticness","Speechiness" using Kmeans.:**')
df2 = df[["Beats Per Minute (BPM)", "Loudness (dB)", "Liveness", "Valence", "Acousticness", "Speechiness"]]
kmeans = KMeans(n_clusters=10)
clusters = kmeans.fit_predict(df2)
df["Music Segments"] = clusters
df["Music Segments"] = df["Music Segments"].map({i: f"Cluster {i+1}" for i in range(10)})
df["Music Segments"].fillna("Cluster 10", inplace=True)
st.dataframe(df)

st.write('**Visualizing:**')
def plot_3d_scatter(df):
    fig = go.Figure()
    for i in df["Music Segments"].unique():
        fig.add_trace(go.Scatter3d(x=df[df["Music Segments"]==i]['Beats Per Minute (BPM)'],
                                   y=df[df["Music Segments"]==i]['Energy'],
                                   z=df[df["Music Segments"]==i]['Danceability'],
                                   mode='markers', marker=dict(size=6, line=dict(width=1)), name=str(i)))
    fig.update_traces(hovertemplate='BPM: %{x} <br>Energy: %{y} <br>Danceability: %{z}')
    fig.update_layout(autosize=True, scene=dict(xaxis_title='BPM', yaxis_title='Energy', zaxis_title='Danceability'))
    return fig
st.plotly_chart(plot_3d_scatter(df))


# Scatter plot
st.write('**Similarity between Top 5 Genres and Clusters:**')
genre_cluster_df = pd.crosstab(df["Top Genre"], df["Music Segments"])

# Filter for top 5 genres
genre_cluster_df = genre_cluster_df.loc[top_genres]

# Prepare data for scatter plot
scatter_data = genre_cluster_df.reset_index().melt(id_vars="Top Genre", var_name="Cluster", value_name="Frequency")
scatter_data["Cluster"] = scatter_data["Cluster"].apply(lambda x: int(x.split()[1]) if x != "Cluster 10" else 10)

# Plot
plt.figure(figsize=(12, 8))
sns.scatterplot(data=scatter_data, x="Cluster", y="Frequency", hue="Top Genre", palette="Set1", s=100)
plt.xlabel("Cluster")
plt.ylabel("Frequency")
plt.legend(title="Top Genre", bbox_to_anchor=(1.05, 1), loc='upper left')
plt.savefig("scatter_plot.png", bbox_inches='tight')  # Save the plot with tight bounding box
st.pyplot(plt)