Scribd
Upload
81 views

Project Amazon Sales Data Analysis

Uploaded by

Somyajit Chakraborty
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
81 views

Project Amazon Sales Data Analysis

Uploaded by

Somyajit Chakraborty
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 12

Project Amazon Sales

Data Analysis
Objective
● Development of a predictive model for predicting sales.
● Perform ETL (Extract-Transform-Load) on dataset.
● Develop dashboard by using tableau.

Benefits
● Better understand and optimise revenue generation in future
● Maximize forecasting accuracy
● Make current sales experience our top priority
Architecture
Data Preprocessing:
● Importing necessary libraries for data analysis such as : Pandas, Numpy, Matplotlib &
Seaborn etc.
● Using pd.read_csv() function stores the data in pandas dataframe named data.
● Using data.column showing columns present in dataframe.
● info() function show basic information of dataframe like null value count of each
column and their data type
● Changing the data type of different column for model training and analysis
● Using describe function on dataframe for getting basic stats of numerical dataset
● Adding extra column to dataframe which contain only month, year and month with
year
● Using isnull().sum() checking out total null value in all the column of dataframe
Exploratory data Analysis
Checking Outliers in the dataframe by using Box Plot

● Box Plot for Total Profit : Here we detect outliers in the specified column using
the Z-score method and found 7 outliers.
● Box Plot of Total Cost : found 5 outliers in Total Cost column

● Box Plot of Total Revenue : Found 6 outliers in Total Revenue column


● Creating a bar chart for Total Revenue and Order Month : where it
showcases the number of order purchased in particular month.

● Calculating the total revenue for each group with respect to Item Type
and then sorting then in descending order.
● Calculating the total profit for each group with respect to Item Type
and then sorting them in descending order.
● Calculating correlation of 'Total Revenue', 'Total Cost' and 'Total Profit'
columns present in dataframe.
Predictive Analytics :
● Label Encoding of Item Type, Sales Channel and Order Priority for model
training.
● Dropping columns Region, Country, Order Date MonthYear, Order ID and
Ship Date.

Pycaret library :
● PyCaret is an open-source, low-code machine learning library in Python.
● Allows users to quickly and easily build, compare, and deploy machine
learning models on structured and tabular data.
● Reduce the amount of code needed to build a model.
● It provides preprocessing and feature engineering functions.
● Automatic model selection and hyperparameter tuning.
● Support for a wide range of machine learning algorithms
● Plotting residuals for Lasso Least Angle Regression based trained model

● Plotting prediction error plot for Lasso Least Angle Regression based trained
model
Implementation of Linear Regression
● Selecting the independent variables and target variable.
● Splitting the data into training and testing datasets.
● Standardizing the dataset.
● Performing fit transform on X_train dataframe.
● Performing fit transform on X_test dataframe.
● Applying Linear Regression on X_train and y_train.
● Calculating mean squared error.
● Creating kernel density estimate plot
● Plotting the predicted values against the actual values to visualize
how well the model is fitting the data.

You might also like