Dalam analisis data, labeling pada plot data clustering dapat membantu dalam mengidentifikasi pattern dan trend yang ada. Berikut adalah 7 gaya labeling yang dapat digunakan:

STYLE 1: LEGEND

Gaya labeling pertama ini menggunakan legenda untuk menunjukkan nama-nama cluster. Legenda ini dapat berupa kotak-kotak atau simbol-simbol yang mewakili setiap cluster.

facet.ax.text(pad, y, 'Distributions of points in clusters:', 
 ha='right', va='bottom', color='black')

for i, label in enumerate(groups.keys()):
 text = facet.ax.text(pad+((i+1)*sep), y, label, 
 ha='right', va='bottom',
 color=customPalette[i])

STYLE 2: LABELS ABOVE CLUSTER

Gaya labeling kedua ini menampilkan nama-nama cluster di atas setiap cluster.

plt.figure(figsize=(5,5))

for i, label in enumerate(groups.keys()):
 plt.scatter(x=data.loc[data['label']==label, 'x'], 
 y=data.loc[data['label']==label,'y'], 
 color=customPalette[i], 
 alpha=0.7)
 
 plt.annotate(label, 
 (data.loc[data['label']==label,['x','y']].mean()[0],
 data.loc[data['label']==label,['x','y']].mean()[1]),
 horizontalalignment='center',
 verticalalignment='center',
 size=20, weight='bold',
 color=customPalette[i])

STYLE 3: LABELS NEXT TO CLUSTER

Gaya labeling ketiga ini menampilkan nama-nama cluster di samping setiap cluster.

labels = {'A': (1.25,1),
 'B': (2.25,4.5),
 'C': (4.75,3.5),
 'D': (4.75,1.5)}

plt.figure(figsize=(5,5))

for i, label in enumerate(groups.keys()):
 plt.scatter(x=data.loc[data['label']==label, 'x'], 
 y=data.loc[data['label']==label,'y'], 
 color=customPalette[i], 
 alpha=0.7)
 
 plt.annotate(label, 
 labels[label],
 horizontalalignment='center',
 verticalalignment='center',
 size=20, weight='bold',
 color=customPalette[i]) 

STYLE 4: LABELS NEXT TO CLUSTER

Gaya labeling keempat ini juga menampilkan nama-nama cluster di samping setiap cluster.

labels = {'A': (1.25,1),
 'B': (2.25,4.5),
 'C': (4.75,3.5),
 'D': (4.75,1.5)}

plt.figure(figsize=(5,5))

for i, label in enumerate(groups.keys()):
 plt.scatter(x=data.loc[data['label']==label, 'x'], 
 y=data.loc[data['label']==label,'y'], 
 color=customPalette[i], 
 alpha=0.7)
 
 plt.annotate(label, 
 labels[label],
 horizontalalignment='center',
 verticalalignment='center',
 size=20, weight='bold',
 color=customPalette[i]) 

STYLE 5: LABELS CENTERED ON CLUSTER MEANS

Gaya labeling kelima ini menampilkan nama-nama cluster di tengah setiap cluster.

plt.figure(figsize=(5,5))

for i, label in enumerate(groups.keys()):
 plt.scatter(x=data.loc[data['label']==label, 'x'], 
 y=data.loc[data['label']==label,'y'], 
 color=customPalette[i], 
 alpha=0.20)
 
 plt.annotate(label, 
 data.loc[data['label']==label,['x','y']].mean(),
 horizontalalignment='center',
 verticalalignment='center',
 size=20, weight='bold',
 color=customPalette[i]) 

STYLE 6: LABELS CENTERED ON CLUSTER MEANS (2)

Gaya labeling keenam ini juga menampilkan nama-nama cluster di tengah setiap cluster.

plt.figure(figsize=(5,5))

for i, label in enumerate(groups.keys()):
 plt.scatter(x=data.loc[data['label']==label, 'x'], 
 y=data.loc[data['label']==label,'y'], 
 color=customPalette[i], 
 alpha=1)
 
 plt.annotate(label, 
 data.loc[data['label']==label,['x','y']].mean(),
 horizontalalignment='center',
 verticalalignment='center',
 size=20, weight='bold',
 color=customPalette[i]) 

STYLE 7: LABELS WITH ICON

Gaya labeling ketujuh ini menampilkan nama-nama cluster dengan simbol-simbol.

plt.figure(figsize=(5,5))

for i, label in enumerate(groups.keys()):
 plt.scatter(x=data.loc[data['label']==label, 'x'], 
 y=data.loc[data['label']==label,'y'], 
 color=customPalette[i], 
 alpha=0.7)
 
 plt.annotate(label + '\n' + icon[label],
 (data.loc[data['label']==label,['x','y']].mean()[0],
 data.loc[data['label']==label,['x','y']].mean()[1]),
 horizontalalignment='center',
 verticalalignment='center',
 size=20, weight='bold',
 color=customPalette[i])

Dengan demikian, dengan menggunakan gaya labeling yang tepat, Anda dapat lebih mudah mengidentifikasi pattern dan trend pada data clustering.