import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
import warnings
warnings.filterwarnings("ignore")

data = pd.read_csv('Mall_Customers.csv')

data.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

Univariate Analysis

The average income of a mall goer is 60,560

Median Income of a person going to the mall is 61,500

The youngest person going to the mall is 18 and the oldest person going to the mall is 70

data.describe()

	CustomerID	Age	Annual Income (k$)	Spending Score (1-100)
count	200.000000	200.000000	200.000000	200.000000
mean	100.500000	38.850000	60.560000	50.200000
std	57.879185	13.969007	26.264721	25.823522
min	1.000000	18.000000	15.000000	1.000000
25%	50.750000	28.750000	41.500000	34.750000
50%	100.500000	36.000000	61.500000	50.000000
75%	150.250000	49.000000	78.000000	73.000000
max	200.000000	70.000000	137.000000	99.000000

We find a normal distribution on this histogram

sns.distplot(data['Annual Income (k$)']);

data.columns

Index(['CustomerID', 'Gender', 'Age', 'Annual Income (k$)',
       'Spending Score (1-100)'],
      dtype='object')

Automated creating histograms using for loop

columns = ['Age', 'Annual Income (k$)', 'Spending Score (1-100)']
for i in columns:
    plt.figure()
    sns.distplot(data[i])

sns.kdeplot(data=data, x='Annual Income (k$)', shade=True, hue=data['Gender']);

For Females age 20-40 their is much more of them going to the mall compared to males

columns = ['Age', 'Annual Income (k$)', 'Spending Score (1-100)']
for i in columns:
    plt.figure()
    sns.kdeplot(data=data, x=i, shade=True, hue=data['Gender']);

Our outlier for the Male income is a person making 138k a year

columns = ['Age', 'Annual Income (k$)', 'Spending Score (1-100)']
for i in columns:
    plt.figure()
    sns.boxplot(data=data, x='Gender', y=data[i], hue=data['Gender']);

56% of our Mall customers are Female

data['Gender'].value_counts(normalize=True)

Gender
Female    0.56
Male      0.44
Name: proportion, dtype: float64

Bivariate Analysis

sns.scatterplot(data=data, x='Annual Income (k$)', y='Spending Score (1-100)')

<Axes: xlabel='Annual Income (k$)', ylabel='Spending Score (1-100)'>

Quick Bivariate Clustering Analysis

#data=data.drop('Spending and Income Cluster', axis=1)
pair = sns.pairplot(data, hue='Gender')
pair.savefig("quick_analysis.png") 

Annual Income of Male is slightly higher, Spending score of females are slighly higher

data.groupby('Gender')[['Age', 'Annual Income (k$)', 'Spending Score (1-100)']].mean()

	Age	Annual Income (k$)	Spending Score (1-100)
Gender
Female	38.098214	59.250000	51.526786
Male	39.806818	62.227273	48.511364

data.corr(numeric_only=True)

	CustomerID	Age	Annual Income (k$)	Spending Score (1-100)
CustomerID	1.000000	-0.026763	0.977548	0.013835
Age	-0.026763	1.000000	-0.012398	-0.327227
Annual Income (k$)	0.977548	-0.012398	1.000000	0.009903
Spending Score (1-100)	0.013835	-0.327227	0.009903	1.000000

Seeing coorelation between two variables using exploratory data analysis

heat = sns.heatmap(data.corr(numeric_only=True), annot=True, cmap='coolwarm')
fig = heat.get_figure()
fig.savefig("heat_map.png") 

Clustering - Univariate, Bivariate, Multivariate

clustering1 = KMeans(n_clusters=3)

clustering1.fit(data[['Annual Income (k$)']])

KMeans(n_clusters=3)

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

clustering1.labels_

array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
       1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
       2, 2])

data['Income Cluster'] = clustering1.labels_
data.head()

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

Majority are in cluster 5

data['Income Cluster'].value_counts()

Income Cluster
0    90
1    74
2    36
Name: count, dtype: int64

clustering1.inertia_

23517.330930930926

inertia_scores=[]
for i in range(1,11):
    kmeans = KMeans(n_clusters = i)
    kmeans.fit(data[['Annual Income (k$)']])
    inertia_scores.append(kmeans.inertia_)

inertia_scores

[137277.28000000003,
 48968.02080832332,
 23528.152173913048,
 15460.42951227089,
 8667.679614837509,
 5443.614973544974,
 4109.451471234081,
 3348.18475968476,
 2584.7879156790923,
 1762.9541125541127]

plt.plot(range(1,11), inertia_scores)

[<matplotlib.lines.Line2D at 0x19a5a9b9e20>]

Cluster 1 has the lowest annual income, Cluster 2 has the highest spending score

data.groupby('Income Cluster')[['Age', 'Annual Income (k$)', 'Spending Score (1-100)']].mean()

	Age	Annual Income (k$)	Spending Score (1-100)
Income Cluster
0	38.722222	67.088889	50.000000
1	39.500000	33.486486	50.229730
2	37.833333	99.888889	50.638889

Bivariate Clustering

clustering2 = KMeans(n_clusters=5)
clustering2.fit(data[['Annual Income (k$)', 'Spending Score (1-100)']])
clustering2.labels_
data['Spending and Income Cluster'] = clustering2.labels_
data.head()

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

inertia_scores2=[]
for i in range(1,11):
    kmeans2 = KMeans(n_clusters = i)
    kmeans2.fit(data[['Annual Income (k$)', 'Spending Score (1-100)']])
    inertia_scores2.append(kmeans2.inertia_)
plt.plot(range(1,11), inertia_scores2)

[<matplotlib.lines.Line2D at 0x19a628ce4e0>]

We find five different clusters in the Annual Income and Spending Score plot

centers = pd.DataFrame(clustering2.cluster_centers_)
centers.columns = ['x','y']

plt.figure(figsize=(10, 8))
plt.scatter(x=centers['x'], y=centers['y'], s=100, c='black', marker='*')
sns.scatterplot(data=data, x = 'Annual Income (k$)', y = 'Spending Score (1-100)', hue='Spending and Income Cluster', palette='tab10')
plt.savefig('clustering_bivariate.png')

Cluster 2 has the highest amount of annual income and spending score which would bring us a lot of money, 53% of them are female

pd.crosstab(data['Spending and Income Cluster'], data['Gender'], normalize='index')

Gender	Female	Male
Spending and Income Cluster
0	0.472222	0.527778
1	0.590909	0.409091
2	0.538462	0.461538
3	0.608696	0.391304
4	0.587500	0.412500

Cluster 1 has the Youngest Age, Lowest Income but a very High Spending score. This could suggest us to make campaigns for young people to spend on Video Game Consoles or Nike Jordans

data.groupby('Spending and Income Cluster')[['Age', 'Annual Income (k$)', 'Spending Score (1-100)']].mean()

	Age	Annual Income (k$)	Spending Score (1-100)
Spending and Income Cluster
0	40.666667	87.750000	17.583333
1	25.272727	25.727273	79.363636
2	32.692308	86.538462	82.128205
3	45.217391	26.304348	20.913043
4	42.937500	55.087500	49.712500

# multivariate clustering
from sklearn.preprocessing import StandardScaler

scale = StandardScaler()

data.head()

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

dff = pd.get_dummies(data, drop_first=True)
dff.head()

	CustomerID	Age	Annual Income (k$)	Spending Score (1-100)	Income Cluster	Spending and Income Cluster	Gender_Male
0	1	19	15	39	1	3	True
1	2	21	15	81	1	1	True
2	3	20	16	6	1	3	False
3	4	23	16	77	1	1	False
4	5	31	17	40	1	3	False

dff.columns

Index(['CustomerID', 'Age', 'Annual Income (k$)', 'Spending Score (1-100)',
       'Income Cluster', 'Spending and Income Cluster', 'Gender_Male'],
      dtype='object')

dff = dff[['Age', 'Annual Income (k$)', 'Spending Score (1-100)', 'Gender_Male']]
dff.head()

	Age	Annual Income (k$)	Spending Score (1-100)	Gender_Male
0	19	15	39	True
1	21	15	81	True
2	20	16	6	False
3	23	16	77	False
4	31	17	40	False

dff = scale.fit_transform(dff)

dff.head()

	0	1	2	3
0	-1.424569	-1.738999	-0.434801	1.128152
1	-1.281035	-1.738999	1.195704	1.128152
2	-1.352802	-1.700830	-1.715913	-0.886405
3	-1.137502	-1.700830	1.040418	-0.886405
4	-0.563369	-1.662660	-0.395980	-0.886405

dff = pd.DataFrame(scale.fit_transform(dff))

inertia_scores3=[]
for i in range(1,11):
    kmeans3 = KMeans(n_clusters = i)
    kmeans3.fit(dff)
    inertia_scores3.append(kmeans3.inertia_)
plt.plot(range(1,11), inertia_scores3)

[<matplotlib.lines.Line2D at 0x19a63ae5310>]

data

	CustomerID	Gender	Age	Annual Income (k$)	Spending Score (1-100)	Income Cluster	Spending and Income Cluster
0	1	Male	19	15	39	1	3
1	2	Male	21	15	81	1	1
2	3	Female	20	16	6	1	3
3	4	Female	23	16	77	1	1
4	5	Female	31	17	40	1	3
...	...	...	...	...	...	...	...
195	196	Female	35	120	79	2	2
196	197	Female	45	126	28	2	0
197	198	Male	32	126	74	2	2
198	199	Male	32	137	18	2	0
199	200	Male	30	137	83	2	2

200 rows × 7 columns

data.to_csv('clustering.csv')