from sklearn.datasets.samples_generator import make_blobs
X, y_true = make_blobs(n_samples=300, centers=3,
                       cluster_std=1.1, random_state=0)
plt.scatter(X[:, 0], X[:, 1], s=50);
from sklearn.cluster import KMeans
kmeans = KMeans(n_clusters=3)
kmeans.fit(X)
y_kmeans = kmeans.predict(X)
plt.scatter(X[:, 0], X[:, 1], c=y_kmeans, s=50, cmap='viridis')

centers = kmeans.cluster_centers_
plt.scatter(centers[:, 0], centers[:, 1], c='black', s=200);