import math
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
from sklearn.cluster import KMeans 
from kneed import KneeLocator
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import silhouette_score

import os
for dirname, _, filenames in os.walk('/kaggle/input'):
    for filename in filenames:
        print(os.path.join(dirname, filename))

wine_df = pd.read_csv("/kaggle/input/wine-dataset-for-clustering/wine-clustering.csv", )
wine_df

sc = StandardScaler()
wine_df_sc = sc.fit_transform(wine_df)

WSS = []

for i in range (1,15):
    kmeans = KMeans(n_clusters=i, n_init=10)
    kmeans.fit(wine_df_sc)
    WSS.append(kmeans.inertia_)

kl = KneeLocator(range(1, 15), WSS, curve="convex", direction="decreasing")
optimal_k = kl.elbow
print("Optimal number of clusters: ", optimal_k)
plt.figure(figsize=(10,6))
plt.title("Elbow Analysis", fontsize=18)
sns.lineplot(x=list(range(1,15)), y=WSS, color="blue", marker="o", markersize=16)
plt.xlabel("Number of clusters")
plt.ylabel("WSS")
plt.vlines(optimal_k, plt.ylim()[0], plt.ylim()[1], linestyles='dashed')
plt.show()

silhouette = []

for i in range(2,15):
    kmeans = KMeans(n_clusters=i, n_init=10)
    cluster_labels = kmeans.fit_predict(wine_df_sc)
    silhouette.append(silhouette_score(wine_df_sc,cluster_labels))   

plt.figure(figsize=(10,6))
plt.title("Silhouette Analysis", fontsize=18)
sns.lineplot(x=list(range(2,15)), y=silhouette, color="green", marker="o", markersize=16)
plt.xlabel("Number of clusters")
plt.ylabel("Silhouette Score")
plt.show()

print("Best Silhouette Clustering is : ", max(silhouette), "with ", silhouette.index(max(silhouette))+2)

kmeans = KMeans(n_clusters=3, n_init=10, random_state=51)
kmeans.fit(wine_df_sc)
wine_df['Cluster']=kmeans.labels_

wine_df.groupby('Cluster').describe(percentiles=[])

from sklearn.decomposition import PCA

data=wine_df.drop("Cluster",axis=1).values
data_sc=sc.fit_transform(data)
pca=PCA(n_components=2)
principal_components = pca.fit_transform(data_sc)
df_pca = pd.DataFrame(principal_components, columns=['PC1', 'PC2'])
df_pca['Cluster']=wine_df['Cluster']

plt.figure(figsize=(10, 7))  
plt.title('Clustering Results using PCA', fontsize=18)
sns.scatterplot(data=df_pca, x='PC1', y='PC2', hue="Cluster", palette="magma", s=100)
plt.grid()
plt.show()

cluster_means=wine_df.groupby('Cluster').mean()
cluster_means