Importing libraries¶
In [38]:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from mpl_toolkits.mplot3d import Axes3D
from sklearn.preprocessing import StandardScaler
Reading CSV¶
In [4]:
data=pd.read_csv("SampleSuperstore.csv")
data
Out[4]:
In [6]:
nRow, nCol = data.shape
print(f'There are {nRow} rows and {nCol} columns')
checking for null values¶
In [7]:
data.isnull().sum()
Out[7]:
Checking the dupilication in data¶
In [8]:
data.duplicated().sum()
Out[8]:
In [9]:
data.nunique()
Out[9]:
Deleting the Variable.¶
In [10]:
col=['Postal Code']
data1=data.drop(columns=col,axis=1)
Correlation Between Variables.¶
In [11]:
data1.corr()
Out[11]:
Covariance of columns¶
In [12]:
data1.cov()
Out[12]:
EDA¶
In [13]:
plt.figure(figsize=(16,8))
plt.bar('Sub-Category','Category', data=data1)
plt.title('Category vs Sub Category')
plt.xlabel('Sub-Catgory')
plt.ylabel('Category')
plt.xticks(rotation=45)
plt.show()
In [14]:
data1.hist(bins=50 ,figsize=(20,15))
plt.show();
Count the total repeatable states¶
In [15]:
data1['State'].value_counts()
Out[15]:
In [16]:
plt.figure(figsize=(15,15))
sns.countplot(x=data1['State'])
plt.xticks(rotation=90)
plt.title("STATE")
plt.show()
In [31]:
sns.set(style="whitegrid")
plt.figure(2, figsize=(20,15))
sns.barplot(x='Sub-Category',y='Profit', data=data, palette='Spectral')
plt.suptitle('Pie Consumption Patterns in the United States', fontsize=16)
plt.show()
In [32]:
figsize=(15,10)
sns.pairplot(data1,hue='Sub-Category')
plt.show
Out[32]:
In [34]:
plt.figure(figsize=(10,4))
sns.lineplot('Discount','Profit', data=data1 , color='y',label='Discount')
plt.legend()
plt.show()
In [36]:
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
In [39]:
state_code = {'Alabama': 'AL','Alaska': 'AK','Arizona': 'AZ','Arkansas': 'AR','California': 'CA','Colorado': 'CO','Connecticut': 'CT','Delaware': 'DE','Florida': 'FL','Georgia': 'GA','Hawaii': 'HI','Idaho': 'ID','Illinois': 'IL','Indiana': 'IN','Iowa': 'IA','Kansas': 'KS','Kentucky': 'KY','Louisiana': 'LA','Maine': 'ME','Maryland': 'MD','Massachusetts': 'MA','Michigan': 'MI','Minnesota': 'MN','Mississippi': 'MS','Missouri': 'MO','Montana': 'MT','Nebraska': 'NE','Nevada': 'NV','New Hampshire': 'NH','New Jersey': 'NJ','New Mexico': 'NM','New York': 'NY','North Carolina': 'NC','North Dakota': 'ND','Ohio': 'OH','Oklahoma': 'OK','Oregon': 'OR','Pennsylvania': 'PA','Rhode Island': 'RI','South Carolina': 'SC','South Dakota': 'SD','Tennessee': 'TN','Texas': 'TX','Utah': 'UT','Vermont': 'VT','Virginia': 'VA','District of Columbia': 'WA','Washington': 'WA','West Virginia': 'WV','Wisconsin': 'WI','Wyoming': 'WY'}
data1['state_code'] = data1.State.apply(lambda x: state_code[x])
In [40]:
state_data = data1[['Sales', 'Profit', 'state_code']].groupby(['state_code']).sum()
fig = go.Figure(data=go.Choropleth(
locations=state_data.index,
z = state_data.Sales,
locationmode = 'USA-states',
colorscale = 'Reds',
colorbar_title = 'Sales in USD',
))
fig.update_layout(
title_text = 'Total State-Wise Sales',
geo_scope='usa',
height=800,
)
fig.show()
Now, let us analyze the sales of a few random states from each profit bracket (high profit, medium profit, low profit, low loss and high loss) and try to observe some crucial trends which might help us in increasing the sales.¶
We have a few questions to answer here.¶
1 What products do the most profit making states buy?¶
2 What products do the loss bearing states buy?¶
3 What product segment needs to be improved in order to drive the profits higher?¶
In [42]:
def state_data_viewer(states):
"""Plots the turnover generated by different product categories and sub-categories for the list of given states.
Args:
states- List of all the states you want the plots for
Returns:
None
"""
product_data = data1.groupby(['State'])
for state in states:
data = product_data.get_group(state).groupby(['Category'])
fig, ax = plt.subplots(1, 3, figsize = (28,5))
fig.suptitle(state, fontsize=14)
ax_index = 0
for cat in ['Furniture', 'Office Supplies', 'Technology']:
cat_data = data.get_group(cat).groupby(['Sub-Category']).sum()
sns.barplot(x = cat_data.Profit, y = cat_data.index, ax = ax[ax_index])
ax[ax_index].set_ylabel(cat)
ax_index +=1
fig.show()
states = ['California', 'Washington', 'Mississippi', 'Arizona', 'Texas']
state_data_viewer(states)
Using Cluster Analysis(K-Mean Clustering)¶
In [43]:
x = data.iloc[:, [9, 10, 11, 12]].values
from sklearn.cluster import KMeans
wcss = []
for i in range(1, 11):
kmeans = KMeans(n_clusters = i, init = 'k-means++',
max_iter = 300, n_init = 10, random_state = 0).fit(x)
wcss.append(kmeans.inertia_)
sns.set_style("whitegrid")
sns.FacetGrid(data, hue ="Sub-Category",height = 6).map(plt.scatter,'Sales','Quantity')
plt.scatter(kmeans.cluster_centers_[:, 0], kmeans.cluster_centers_[:,1],
s = 100, c = 'yellow', label = 'Centroids')
plt.legend()
plt.show()
In [44]:
sns.set_style("whitegrid")
sns.FacetGrid(data, hue ="Sub-Category",height = 6).map(plt.scatter,'Sales','Profit')
plt.scatter(kmeans.cluster_centers_[:, 0], kmeans.cluster_centers_[:,1],
s = 100, c = 'yellow', label = 'Centroids')
plt.legend()
plt.show()
In [45]:
fig, ax = plt.subplots(figsize = (10 , 6))
ax.scatter(data1["Sales"] , data1["Profit"])
ax.set_xlabel('Sales')
ax.set_ylabel('Profit')
ax.set_title('Sales vs Profit')
plt.show()
From the Above data Visualization and Clustering we can see that in Which states and in which Category Sales and profits are High or less,We can improve in that States By Providing Discounts in prefered Range so that Company and cosumer both will be in profit.So For Deciding that Range we have to do some Technical Analysis.One can Do it through Factor Analysis,or also can Do it throgh neural networks.¶
In [ ]: