# Packages
import pandas as pd
import numpy as np

import seaborn as sns
sns.set_style("dark")

import matplotlib.pyplot as plt
import sys
if not sys.warnoptions:
    import warnings
    warnings.simplefilter("ignore")

# Importing data 
df = pd.read_csv('./Mall_Customers.csv')

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 200 entries, 0 to 199
Data columns (total 5 columns):
 #   Column                  Non-Null Count  Dtype 
---  ------                  --------------  ----- 
 0   CustomerID              200 non-null    int64 
 1   Gender                  200 non-null    object
 2   Age                     200 non-null    int64 
 3   Annual Income (k$)      200 non-null    int64 
 4   Spending Score (1-100)  200 non-null    int64 
dtypes: int64(4), object(1)
memory usage: 7.9+ KB

df.head()

	CustomerID	Gender	Age	Annual Income (k$)	Spending Score (1-100)
0	1	Male	19	15	39
1	2	Male	21	15	81
2	3	Female	20	16	6
3	4	Female	23	16	77
4	5	Female	31	17	40

---

Average annual income per gender

# renaming column names for convenience 
df = df.rename(columns={'Annual Income (k$)':'annual_income','Spending Score (1-100)':'spending_score'})
df

	CustomerID	Gender	Age	annual_income	spending_score
0	1	Male	19	15	39
1	2	Male	21	15	81
2	3	Female	20	16	6
3	4	Female	23	16	77
4	5	Female	31	17	40
...	...	...	...	...	...
195	196	Female	35	120	79
196	197	Female	45	126	28
197	198	Male	32	126	74
198	199	Male	32	137	18
199	200	Male	30	137	83

200 rows × 5 columns

g1 = df.groupby('Gender')[['annual_income','spending_score']].mean()
g1.plot(kind='barh',figsize=(14,6),title='Average annual income and spending score per gender')

<AxesSubplot:title={'center':'Average Annual_income and spending score per gender'}, ylabel='Gender'>

# Correlation
df.corr()

	CustomerID	Age	annual_income	spending_score
CustomerID	1.000000	-0.026763	0.977548	0.013835
Age	-0.026763	1.000000	-0.012398	-0.327227
annual_income	0.977548	-0.012398	1.000000	0.009903
spending_score	0.013835	-0.327227	0.009903	1.000000

Income distribution per gender

plt.figure(figsize=(14,6))
sns.histplot(x=df["annual_income"],hue=df['Gender'],kde=True,multiple='stack').set(title='Distrbution of income per Gender')

[Text(0.5, 1.0, 'Distrbution of income per Gender')]

Average annual income per age group

conditions = [
    (df['Age']<=21),
    (df['Age']>21) & (df['Age']<=40),
    (df['Age']>40) 
]

values = ['<17','17-40','>40']
df['age_group'] = np.select(conditions,values)
df.groupby('age_group')['annual_income'].mean().plot(kind='barh',title='Average annual income per age group',color=['teal','orange','green'],figsize=(10,6))

<AxesSubplot:title={'center':'Average annual income per age group'}, ylabel='age_group'>

---

K Means Clustering

df2 = df[['spending_score','annual_income']]
df2.head()
plt.figure(figsize=(14,6))
sns.scatterplot(df2['spending_score'],df2['annual_income'],y='Annual Income',x='spending_score')

<AxesSubplot:xlabel='spending_score', ylabel='annual_income'>

# kmean
from sklearn.cluster import KMeans

The ‘Within Sum of squares’ or WSS value is defined as the sum of the squared distance between each member of the cluster and its centroid. Mathematically, this can be represented as: WSS = SUM(X-C)^2

wss=[]
for i in range(1,11):
    km=KMeans(n_clusters=i)
    km.fit(df2)
    wss.append(km.inertia_)
    
plt.figure(figsize=(16,6))
plt.plot(wss,marker="o",c='r')
plt.xlabel('Number of k (Clusters)')
plt.ylabel('WSS')
plt.show()

Selecting 5 clusters based on the elbow graph above

km=KMeans(n_clusters=5)
km.fit(df2)
cluster=km.predict(df2)
df2['Group'] = cluster
df2.head()

	spending_score	annual_income	Group
0	39	15	3
1	81	15	4
2	6	16	3
3	77	16	4
4	40	17	3

Plotting

# Average annual income 
df['annual_income'].mean()

60.56

# Average spending income 
df['spending_score'].mean()

50.2

plt.figure(figsize=(16,6))
sns.scatterplot(df2['spending_score'],df2['annual_income'],y='Annual Income',x='spending_score',hue=df2['Group'],s=100,palette=['teal','orange','salmon','olivedrab','peru',]).set(title='Annual_incoming/Spending score')

plt.axvline(50.2, color='grey',linestyle ="--") # vertical
plt.axhline(60.56, color='grey',linestyle ="--") # horizontal

<matplotlib.lines.Line2D at 0x203ae5a7d90>

from prettytable import PrettyTable

# columns
myTable = PrettyTable(["Cluster", "Description"])
 
# Add rows
myTable.add_row(["0", "Average spending score with Average annual income"])
myTable.add_row(["1", "Lower than average spending socre with high annual income"])
myTable.add_row(["2", "High spending score with higher than average annual income"])
myTable.add_row(["3", "Lower than average spending score with low annual income"])
myTable.add_row(["4", "High spending score with lower than average annual income"])

myTable

Cluster	Description
0	Average spending score with Average annual income
1	Lower than average spending socre with high annual income
2	High spending score with higher than average annual income
3	Lower than average spending score with low annual income
4	High spending score with lower than average annual income