Project 2

1. Importing Dataset:

In[1]:

# 'Importing the required Libraries'
from os import execl
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans
import seaborn as sns

# 'Importing the dataset using pandas'
df=pd.read_csv("auto-mpg_kaggle.csv")

# 'Printing the first 5 rows in the Data Frame'
df.head()

Out[1]:

	mpg	cylinders	displacement	horsepower	weight	acceleration	model year	origin	car name
0	18.0	8	307.0	130	3504	12.0	70	1	chevrolet chevelle malibu
1	15.0	8	350.0	165	3693	11.5	70	1	buick skylark 320
2	18.0	8	318.0	150	3436	11.0	70	1	plymouth satellite
3	16.0	8	304.0	150	3433	12.0	70	1	amc rebel sst
4	17.0	8	302.0	140	3449	10.5	70	1	ford torino

In[2]:

# "Describing the Data frame"
df.describe()
print(df.dtypes)

Out[2]:

	mpg	cylinders	displacement	horsepower	weight	acceleration	model year	origin
count	398.000000	398.000000	398.000000	398.000000	398.000000	398.000000	398.000000	398.000000
mean	23.514573	5.454774	193.425879	104.185930	2970.424623	15.568090	76.010050	1.572864
std	7.815984	1.701004	104.269838	38.292729	846.841774	2.757689	3.697627	0.802055
min	9.000000	3.000000	68.000000	46.000000	1613.000000	8.000000	70.000000	1.000000
25%	17.500000	4.000000	104.250000	75.000000	2223.750000	13.825000	73.000000	1.000000
50%	23.000000	4.000000	148.500000	93.500000	2803.500000	15.500000	76.000000	1.000000
75%	29.000000	8.000000	262.000000	125.000000	3608.000000	17.175000	79.000000	2.000000
max	46.600000	8.000000	455.000000	230.000000	5140.000000	24.800000	82.000000	3.000000

2. EDA: Finding Correlation between Parameters:

In[3]:

# 'Finding the Correlation between the input parameters to predict the output'
corr = df.corr()
corr.style.background_gradient(cmap='PuOr')

Out[3]:

In[4]:

# 'Heatsink correlation'
sns.heatmap(df.corr(), annot = True)
plt.savefig("auto-mpg_kaggle.jpg")
plt.show()

Out[4]:

From the above two correlation technique it is clear that the following parameters will have more weight in predicting the perfect clusters:

1. No.of cylinders.

2. Displacement.

3. Horse Power.

4. Weight.

In[5]:

# 'Pair plot'
sns.pairplot(df, diag_kind = "kde", markers = "+")
plt.show()

Out[5]:

In[6]:

# 'Dropping the colums which are not useful'
df_1 = df.drop(columns=["model year", "car name"])
df_1.head()

Out[6]:

	mpg	cylinders	displacement	horsepower	weight	acceleration	origin
0	18.0	8	307.0	130	3504	12.0	1
1	15.0	8	350.0	165	3693	11.5	1
2	18.0	8	318.0	150	3436	11.0	1
3	16.0	8	304.0	150	3433	12.0	1
4	17.0	8	302.0	140	3449	10.5	1

In[7]:

# 'Implementing the all the colums in standard format'
scaler = StandardScaler()
df_scaled = scaler.fit_transform(df_1)
df_scaled

Out[7]:

array([[-0.7064387 , 1.49819126, 1.0906037 , ..., 0.63086987, -1.29549834, -0.71514478],

[-1.09075062, 1.49819126, 1.5035143 , ..., 0.85433297, -1.47703779, -0.71514478],

[-0.7064387 , 1.49819126, 1.19623199, ..., 0.55047045, -1.65857724, -0.71514478],

...,

[ 1.08701694, -0.85632057, -0.56103873, ..., -0.79858454, -1.4407299 , -0.71514478],

[ 0.57460104, -0.85632057, -0.70507731, ..., -0.40841088, 1.10082237, -0.71514478],

[ 0.95891297, -0.85632057, -0.71467988, ..., -0.29608816, 1.39128549, -0.71514478]])

3. Clustering:

3a.K-Means Clustering:

To reduce the feature size using the PCA technique.

In[8]:

# Reducing the Number of Features in the Dataset using PCA
pca = PCA(2)
pca_data = pca.fit_transform(df_scaled)

plt.rcParams["figure.figsize"] = (12, 8)
plt.rcParams["figure.dpi"] = 80
plt.scatter(x=pca_data[:,0], y=pca_data[:,1], lw=2)
plt.xlabel("Principle Component 1")
plt.ylabel("Principle Component 2")
plt.title("Principle Components Analysis")
plt.show()

Out[8]:

In[9]:

# 'To find the optimum clusters by Elbow Curve method'
k_range=range(1,10)
sse=[]
for k in k_range:
    km=KMeans(n_clusters=k)
    km.fit(df_scaled)
    sse.append(km.inertia_)
print(sse)
plt.plot(k_range,sse)
plt.show()

Out[9]:

[2786.0000000000014, 1242.5014943571473, 844.2488691571539, 670.6355481197716, 567.9699640996866, 502.44711077200765, 445.74932889207753, 408.4932093481778, 384.2047906386553]

From the above elbow curve the optimum no.of clusters will be 3.

In[10]:

# 'K-Means Clustering'
kmeans = KMeans(n_clusters = 3, init = 'k-means++', random_state = 42)
y_kmeans = pd.Series(kmeans.fit_predict(df_scaled))
y_kmeans

centroids = kmeans.cluster_centers_
centroids_pca = pca.transform(centroids)

pca_dataframe = pd.DataFrame(pca_data, columns=["PCA 1", "PCA 2"])
pca_dataframe["Cluster"] = y_kmeans
pca_dataframe.sample(10)

plt.rcParams["figure.figsize"] = (12, 8)
plt.rcParams["figure.dpi"] = 80
sns.scatterplot(x="PCA 1", y="PCA 2", hue="Cluster", data=pca_dataframe, palette=['green','orange','brown'], s=100)
plt.scatter(x=centroids_pca[:, 0], y=centroids_pca[:, 1], marker="+", s=500, linewidths=3, lw=4,, zorder=10)
plt.title("Visualizing Clusters")
plt.show()

Out[10]:

In[11]:

# 'Pair plot after Clustering'
df["Cluster"] = y_kmeans

sns.pairplot(df, hue="Cluster", palette=['green','orange','brown'])
plt.show()

Out[11]:

3b.Hierarchical clustering:

In[12]:

# 'Hierarchical clustering'
import scipy.cluster.hierarchy as sc
from scipy.cluster.hierarchy import dendrogram , linkage
 
#Linkage Matrix
Z = linkage(df_1, method = 'ward')
 
#plotting dendrogram
dendro = dendrogram(Z)
plt.title('Dendrogram')
plt.ylabel('Euclidean distance')
plt.show()

Out[12]: