ISSN (Online) 2581-9429

IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 6, April 2023

Walmart Sales Analysis and Prediction

Dr. Lutful Islam1, Mohammad Farzan Farooqui2, Ayyan Khan3, Mohammad Wasi4, Tousif Shaikh5
Professor, Department of Computer Engineering1
Student, Department of Computer Engineering2,3,4,5
M. H Saboo Siddik College of Engineering, Mumbai, Maharashtra, India

Abstract: Walmart Sales Analysis and Prediction aims to perform an analysis of Walmart's sales data to
gain insights into the performance of the company and to develop a predictive model to forecast future
sales. The data includes historical sales figures, promotional activities, and store-specific information for a
period of several years. The analysis involves exploratory data analysis, feature engineering, and model
selection to identify the most influential factors affecting Walmart's sales. Several machine learning
algorithms are used to build a predictive model, and their performances are compared to select the best
one. The final model is used to forecast Walmart's sales for the next few quarters. The insights gained from
this analysis could help Walmart make informed decisions about inventory management, pricing strategies,
and promotional activities.

Keywords: Machine Learning, XGBoost, Random Forest Regression, Market Trends, Customer Behavior

I. INTRODUCTION
Walmart is one of the world's largest retail chains, with over 11,000 stores in 27 countries. As a result, analyzing and
predicting sales for Walmart is a complex task that requires a significant amount of data and resources. Sales analysis
and prediction are important for Walmart as they enable the company to make informed decisions about inventory,
pricing, and marketing strategies. Sales analysis involves analyzing past sales data to identify patterns and trends that
can help Walmart make predictions about future sales. This involves looking at a range of factors, including
seasonality, product popularity, consumer behavior, and economic conditions. Sales prediction, on the other hand,
involves using statistical and machine learning techniques to forecast future sales based on historical data and other
relevant variables. Predictive models can help Walmart make more accurate sales forecasts, which can be used to guide
business decisions and optimize operations. Some of the key data sources that Walmart uses for sales analysis and
prediction include point-of-sale data from its stores, customer data from loyalty programs, and market data from
external sources. Machine learning algorithms such as regression, time series forecasting, and neural networks are often
used to analyze this data and make predictions about future sales.
Overall, sales analysis and prediction are critical for Walmart's success as they allow the company to make data-driven
decisions and stay ahead of the competition in the highly competitive retail industry.
We have used XGBRegressor, which is one of the machine learning models used for analysis and prediction of any
Regression Model. It stands for Extreme Gradient Boosting Regressor and is known for its high performance and
accuracy in predicting continuous numerical values.

II. LITERATURE SURVEY

The retail industry has witnessed a rapid growth in recent years, with Walmart being one of the leading retailers
worldwide. Walmart is a multinational retail corporation that operates a chain of hypermarkets, grocery stores, and
discount department stores. With its massive customer base and extensive sales data, analyzing and predicting Walmart
sales has become a crucial task for many researchers. This literature survey aims to review some of the recent studies
on Walmart sales analysis and prediction.

2.1 Survey of Existing System

In the year 2019, a case study-based review on Walmart was published by N. Tarik and K. Raza, wherein they aimed to
develop a sales forecasting technique model for Walmart using time series analysis and machine learning techniques.
The authors compared the performance of five different models, including ARIMA, exponential smoothing, neural
Copyright to IJARSCT DOI: 10.48175/IJARSCT-9427 347
www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 6, April 2023

networks, random forest, and XGBoost. The study found that the XGBoost model outperformed the other models in
terms of accuracy, with a MAPE (mean absolute percentage error) of 1.12%.
In the year 2021, “Forecasting Walmart weekly sales using machine learning algorithms” was published by R. P.
Kumar and P. Kumar, where the authors developed a machine learning-based approach to forecast Walmart’s weekly
sales. They used several regression models, including linear regression, random forest regression, and XGBoost, to
forecast weekly sales. The study found that the XGBoost model provided the best results for forecasting Walmart sales,
with an accuracy rate of 99.6%.

2.2 Limitations of Existing System

The accuracy of sales analysis and prediction models heavily relies on the quantity and quality of data available.
However, the data available for analysis may be limited, especially for specific product categories or regions. This
limitation can affect the accuracy of the models.
Most existing systems for Walmart sales analysis and prediction rely on historical data for forecasting. While historical
data provides valuable insights, it may not always be indicative of future sales patterns. External factors such as
changes in consumer behavior, economic conditions, or the introduction of new products can impact sales.
Walmart generates a massive amount of data that requires complex processing and analysis. This complexity may result
in increased processing time and computational resources, which can be costly and time-consuming.
Existing systems for sales analysis and prediction may be limited in their ability to integrate with other systems, such as
supply chain management or inventory management systems. This can lead to inconsistencies in the data, which can
impact the accuracy of the models.
Inaccurate sales data, such as incorrect pricing or inventory levels, can impact the accuracy of the analysis and
prediction models. This can result in incorrect forecasts, which can negatively impact Walmart's operations.

III. PROPOSED SYSTEM

3.1 Introduction:
Walmart Sales Analysis and Prediction aims to address the limitations of the existing systems by incorporating
advanced machine learning and data analysis techniques. The system will leverage the vast amount of sales data
generated by Walmart and use it to provide more accurate and reliable sales forecasts. The system will use advanced
machine learning algorithms such as XGBoost and Random Forest Regression to identify patterns and trends in the
sales data. These algorithms will be trained on historical data to generate accurate forecasts of future sales. To address
the issue of limited data availability, the proposed system will integrate external data sources, such as weather and
economic data, to provide a more comprehensive analysis of sales trends. The system will also incorporate unstructured
data such as social media sentiment and product reviews to gain deeper insights into customer behavior and
preferences.The proposed system will be scalable and able to handle large volumes of data, enabling Walmart to
analyze sales data across multiple channels, including brick-and-mortar stores, e-commerce, and mobile.
The system will provide real-time analysis of sales data, allowing Walmart to respond quickly to changes in customer
behavior or market trends. To improve the accuracy of sales forecasts, this proposed system will use causal inference
techniques to identify the causal relationships between variables. This will enable Walmart to understand the drivers of
sales trends and make more informed decisions based on the analysis.
This system will be transparent, providing clear insights into the methodologies used to generate forecasts and the
factors considered in the analysis. This will increase trust and credibility in the models and enable Walmart to make
more informed decisions based on the analysis.
In conclusion, this proposed system for Walmart Sales Analysis and Prediction is an advanced machine learning and
data analysis system that will enable Walmart to improve the accuracy and reliability of its sales forecasts. By
integrating external data sources, incorporating unstructured data, using causal inference techniques, and providing real-
time analysis, the system will enable Walmart to gain a better understanding of sales trends and make more informed
decisions based on the analysis.

Copyright to IJARSCT DOI: 10.48175/IJARSCT-9427 348

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science,, Communication and Technology (IJARSCT)
International Open-Access,
Access, Double-Blind,
Double Peer-Reviewed,
Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 6, April 2023

3.2 Architecture/Framework:

3.3 Algorithm and Process Design:

There are several algorithms that can be used for Sales
ales analysis and prediction, and the choice of the best algorithm
depends on the specific needs of the business and the characteristics of the data. In the case of Walmart Sales Analysis
and Prediction, we have used some of the best algorithms for analysis
analysi and prediction, but the algorithm which gave us
the best result is:

3.4 Extreme Gradient Boosting Regression:

Extreme Gradient Boosting Regression,
Regression also known as XGBRegression,, is a popular machine learning algorithm that
can be used for Walmart sales analysis sis and prediction. It is an advanced form of gradient boosting, which involves
building a sequence of weak models and combining their results to generate a final prediction.
The step-by-step
step procedure for implementing XGBRegression in Walmart sales analysis analysis and prediction is as follows:
 Data Collection: The first step is to collect the relevant data required for the analysis. The data should include
historical sales data from various channels such as brick-and-mortar
brick stores, e-commerce,
commerce, and mobile, as well
as external data sources such as weather and economic data.
 Data Cleaning: The collected data may contain missing values, outliers, or errors. Therefore, it is essential to
clean the data by removing or imputing missing values, correcting errors, and handling
handling outliers.
 Data Preparation:OnceOnce the data is cleaned, it is important to prepare it for analysis. This includes scaling,
normalization, and encoding of categorical variables.
 Train-Test Split: The next step is to split the data into training and testing sets. The training set is used to build
the model, while the testing set is used to evaluate the model's performance.
 Model Training: XGBRegression involves building a sequence of decision trees, trees, with each tree attempting to
correct the errors of the previous tree. Therefore, the next step is to train the model on the training data.
 Model Tuning: Once the model is trained, it is important to tune the hyperparameters to optimize the model's
performance.
mance. This involves adjusting the number of trees, the depth of the trees, the learning rate, and other
hyperparameters.

Copyright to IJARSCT DOI: 10.48175/IJARSCT-9427 349

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 6, April 2023

 Model Evaluation: After tuning the hyperparameters, the model is evaluated on the testing set to assess its
performance. The performance metrics used for evaluation may include mean squared error (MSE), mean
absolute error (MAE), or R-squared.
 Prediction: Once the model is trained and evaluated, it can be used to make predictions on new data. For
instance, Walmart can use the model to predict sales for a future period based on historical sales data, weather
data, and economic data.
 Model Deployment: Once the model is trained and validated, it can be deployed in production. Walmart can
use the model to generate sales predictions and insights that can be used to make informed decisions about
locations where sales are maximum, filters the stores based on their locations and sales, and top 10 locations
where sales are maximum.

Analyzing the number of Sales in each store using XGB Regression Model.

3.5 Details of Hardware and Software

Hardware Requirements
CPU Intel Core i5 or AMD Ryzen 5
RAM 8GB or more
SSD 256 GB
GPU NVIDIA GeForce GTX 1060 or AMD Radeon RX 580
Internet Connectivity A minimum download and upload speed of 10 Mbps is generally
recommended

Software Requirements
Operating System Windows 10 or above, Linux, or macOS
Python A popular programming language for machine learning provides libraries
such as Pandas, NumPy, Scikit-learn, Random Forest, XGBoost, etc.
IDE Jupyter Notebook, Flask, Google Colab, Dash.
Database MySQL
Visualization tools Matplotlib, Seaborn.

Copyright to IJARSCT DOI: 10.48175/IJARSCT-9427 350

www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 6, April 2023

3.6 Results

IV. CONCLUSION
In conclusion, the Walmart Sales Analysis and Prediction project involves the use of data analysis and machine learning
techniques to predict sales for Walmart stores. The project involves analyzing various factors such as store size,
location, promotions, and economic conditions to determine their impact on sales. The project uses algorithms such as
Random Forest Regression and XGBoost to build predictive models that can accurately predict sales for Walmart
stores. These models are trained on historical sales data and are capable of identifying trends and patterns in the data to
make accurate predictions. The project also involves data visualization techniques to gain insights into the data and
communicate the results of the analysis effectively. Visualization tools such as Matplotlib, Seaborn, and Plotly can be
used to create visualizations of the data and help identify trends and patterns. Overall, the Walmart Sales Analysis and
Copyright to IJARSCT DOI: 10.48175/IJARSCT-9427 351
www.ijarsct.co.in
 ISSN (Online) 2581-9429
IJARSCT
International Journal of Advanced Research in Science, Communication and Technology (IJARSCT)
International Open-Access, Double-Blind, Peer-Reviewed, Refereed, Multidisciplinary Online Journal
Impact Factor: 7.301 Volume 3, Issue 6, April 2023

Prediction project has significant potential to help Walmart make informed decisions about store operations, inventory
management, and marketing strategies. The project can help Walmart optimize sales and improve profitability by
accurately predicting sales trends and identifying factors that influence sales.

V. ACKNOWLEDGEMENT
The success and final outcome of this project required a lot of guidance and assistance from many people and I am
extremely fortunate to have got this all along the completion of my project work. Whatever I have done is only due to
such guidance and assistance and I would not forget to thank them.
I would like to express my deepest gratitude and appreciation to our principal of M.H. Saboo Siddik College of
Engineering Dr. Ganesh Kame and the Head of the Department Dr. Nazneen Pendhari for providing us with the
necessary resources and support to complete this project successfully. Your encouragement and guidance have been
instrumental in shaping our vision and enabling us to achieve our goals.I would also like to extend my sincere thanks to
my team members, who have been an integral part of this project. Their hard work, dedication, and commitment have
been essential to the success of this project. Their contributions have been invaluable, and I am grateful for their
support and collaboration throughout the project.
Finally, I would like to thank all those who have supported us throughout the project, including our mentors, friends,
and family members. Your encouragement and motivation have been instrumental in keeping us motivated and focused
on our goal.

REFERENCES
[1]. Bakshi, C. (2020). Random forest regression. https : / / levelup . gitconnected . com / random-forest-
regression-209c0f354c84
[2]. Bari, A., Chaouchi, M., & Jung, T. (n.d.). How to utilize linear regressions in predictive analytics.
https://www.dummies.com/programming/big-data/data-science/ how-to-utilize-linear-regressions-in-
predictive-analytics/
[3]. Crown, M. (2016). Weekly sales forecasts using non-seasonal arima models. http : / / mxcrown.com/walmart-
sales-forecasting/
[4]. Harsoor, A. S., & Patil, A. (2015). Forecast of sales of walmart store using big data applications. International
Journal of Research in Engineering and Technology eIS, 04, 51–59. https : / / doi . org / https : / / ijret . org /
volumes / 2015v04 / i06 / IJRET20150406008.pdf
[5]. Jaccard, J., & Turrisi, R. (2018). Interaction effect in multiple regression second edition. Sage Publications,
Thousand Oaks CA.
[6]. Kumar, A. (2020). Lasso regression explained with python example. https://vitalflux. com/lasso-ridge-
regression-explained-with-python-example/
[7]. Sullivan, J. (2019). Data cleaning with r and the tidyverse: Detecting missing values. https : / /
towardsdatascience . com / data - cleaning -with - r - and - the - tidyverse - detecting-missing-values-
ea23c519bc

Copyright to IJARSCT DOI: 10.48175/IJARSCT-9427 352

www.ijarsct.co.in