VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY UNIVERSITY OF INFORMATION TECHNOLOGY ADVANCED PROGRAM IN INFORMATION SYSTEM TA THI KIM BÌNH-18520510 TRIEU KIM NGAN-18521137 GRADUATION THESIS ANALYSIS AND FORECASTING THE UNIT SALES OF WALMART RETAIL GOODS BACHELOR OF ENGINEERING IN INFORMATION SYSTEMS HO CHI MINH CITY, 2022 VIETNAM NATIONAL UNIVERSITY HO CHI MINH CITY UNIVERSITY OF INFORMATION TECHNOLOGY ADVANCED PROGRAM IN INFORMATION SYSTEM TA THI KIM BÌNH-18520510 TRIEU KIM NGAN-18521137 GRADUATION THESIS ANALYSIS AND FORECASTING THE UNIT SALES OF WALMART RETAIL GOODS BACHELOR OF ENGINEERING IN INFORMATION SYSTEMS THESIS ADVISOR PHD. DO TRONG HOP HO CHI MINH CITY, 2022 ASSESSMENT COMMITTEE The Assessment Committee is established under the Decision. by the Rector of the University of Information Technology. —= eeee cece eee ees — Chairman.
THANK YOU With all gratitude, first of all, we would like to express my deep gratitude to our teacher Dr. Do Trong Hop for guiding and helping us in the process of completing our graduation thesis. We would also like to thank the teachers who have dedicatedly conveyed important and useful knowledge that is not only the foundation for the process of making the thesis but also a preparation for the road ahead. In addition, we would also like to thank the Board of Directors of the University of Information Technology as well as the departments of the Faculty of Information Systems of the school for creating conditions and facilities for us to have the opportunity and learning environment.
Due to limited knowledge and reasoning ability, the thesis still has certain shortcomings. We look forward to receiving the contributions of teachers and teachers to make my graduation thesis more complete. Finally, we would like to wish the teachers of the Board of Directors and functional departments of the University of Information Technology good health and success in their careers. Thank you sincerely! UNIVERSITY OF INFORMATION TECHNOLOGY Advanced ADVANCED PROGRAM Education IN INFORMATION SYSTEMS Program THESIS PROPOSAL THESIS TITLE: Analysis and Forecasting the unit sales of Walmart retail goods Advisor: Dr.
Đỗ Trọng Hop Duration: (From 70/9/2022 to 15/1/2022) Students: (Ta Thi Kim Binh -18520510 Triệu Kim Ngân - 18521137) Contents: (Describe the details of what to be done, scope, objectives, methodologies, expected results) 1. Scope *Data analyst *Building prediction model 2. Objectives Forecasting by identifying the method(s) that provide the most accurate point forecasts for each of the 42,840-time series. It uses grouped unit sales data, starting at the product- store level and being aggregated to that of product departments, product categories, stores, and three geographical areas: the States of California (CA), Texas (TX), and Wisconsin (WI).
Besides the time series data, it includes explanatory variables such as sell prices, promotions, days of the week, and special events (e. Super Bowl, Valentine’s Day, and Orthodox Easter) that typically affect unit sales and could improve forecasting accuracy. *Dataset: The dataset, generously made available by Walmart, involves the unit sales of various products sold in the USA. More specifically, the dataset involves the unit sales of 3,049 products, classified in 3 product categories (Hobbies, Foods, and Household) and 7 product departments, in which the above-mentioned categories are disaggregated.
The products are sold across ten stores, located in three States (CA, TX, and WI). The dataset consists of the following three (3) files: File 1: “calendar.csy” Contains information about the dates the products are sold. date: The date in a “y-m-d” format. wm_yr_wk: The id of the week the date belongs to.
weekday: The type of the day (Saturday, Sunday,. The id of the weekday, starting from Saturday. month: The month of the date. year: The year of the date.
event_name_1: If the date includes an event, the name of this event. event_type_1: If the date includes an event, the type of this event. event_name_2: If the date includes a second event, the name of this event. event_type_2: If the date includes a second event, the type of this event.
snap_CA, snap_TX, and snap_WI: A binary variable (0 or 1) indicating whether the stores of CA, TX or WI allow SNAP purchases on the examined date. | indicates that SNAP purchases are allowed. File 2: “sell_prices.csv” Contains information about the price of the products sold per store and date. store_id: The id of the store where the product is sold.
item_id: The id of the product. wm_yr_wk: The id ofthe week. sell_price: The price of the product for the given week/store. The price is provided per week (average across seven days).
If not available, this means that the product was not sold during the examined week. File 3: “sales_train.csv” Contains the historical daily unit sales data per product and store. item_id: The id of the product. dept_id: The id of the department the product belongs to.
cat_id: The id of the category the product belongs to. store_id: The id of the store where the product is sold. state_id: The State where the store is located. d_1941: The number of units sold at day i, starting from 2011-01- 29.
Methodologies In this thesis, we intend to determine the elements that have an impact on the retail shop business by looking at the sales data of Walmart stores that are geographically located in different places and performing a sales projection for the following 28 days.We applied EDA (Exploratory Data Analysis) techniques to analyze influential variables and ANOVA (Analysis of variance) statistical methods to retest variables. Then, we built two forecasting models ARIMA (Autoregressive Integrated Moving Average) and LSTM (Long Short-Term Memory) to find the best model. After we calculate RMSE (Root-mean-square error) metric on two models to compare two forecasting models for financial time series. Expected results Our results show that the LSTM model achieves the lowest error metrics than the ARIMA model of the tested models.
With an RMSE of 0.643, LSTM succeeds in outperforming than the RMSE of ARIMA 0. The LSTM model was applied across the board to anticipate sales units over the next 28 days. Research timelines: (Plan of action: describe the planning and the assigned tasks for each student) Duration: 5 months Task Ta Thi Kim Binh Trieu Kim Ngan Survey the problem and get x x an overview Research methods and k x approaches to problem solving Deploy the experimental x x model Evaluate and analyze x x results Write reports and reports x x Approved by the advisor(s) Ho Chi Minh city,. Signature(s) of advisor(s) Signature(s) of student(s) TABLE OF CONTENTS H\ vo.
2 Problem sfaf€Tmer[. LH TH HH HH TH. Forecasting Indonesia exports using a hybrid model ARIMA-LSTM.2 Time series forecasting the financial budget using ARIMA and LSTM Model s. Profit prediction using ARIMA and LSTM models in time series forecasting Tố ốố ốố ốe.1 Process OV€TVI€W.
St TT HH HH”. tt TH HH Hit 8 Chapter 2. -- - St ST E1 2 E1 HH1 0g H1 001.1 Exploratory data analysis (ElDA)).2 ANOVA (Analysis of Variance) testing method.3 Measures of forecast error: RMSE .5 Deep learning an. Experiment and Discussion .2 Approach analysis dafaSef.1 Information of device.3 Data analysis results.1 Units sold for each product category according years.2 Units sold for each product category according months.3 Units sold for each product category according weekday.4 Units sold for each product category according event .5 Units sold for each product category according stOre.6 Units sold for each product category according sfate.4 Building forecasting model reSuÏfS.--- ¿+ 5252 2 £++ke£vrtzxerrkererxee 48 Chapter 4.
Conclusion and Future WOFÍK. -- - 5+ 5t+t+Et+v+Evxrterrrrxrrerxerrrrrerrrrre 52 References NA. 54 LIST OF FIGURES Figure 1. 2 Forecasting DTOC€SS.- -- + nền như 8 Figure 2.
I The fundamental steps of the exploratory data analysis process. 2 Residuals on a scatter pÏOf. - - ¿+ St EE*Ek£kEEEEkEEEEEEkEEEEkrkrrkrkee 13 Figure 2. 3 The machine learning process works.
4 Flow Chart of ARIMA Model. 5 A deep neural network consists of an input layer, multiple hidden layers and an output layer, all consisting of nodes. 6 A node ina hidden layer of a deep neural network takes an input, performs a calculation, and passes on an output to nodes in the next layer. 7 Each neural node in a deep neural network performs a calculation by taking inputs from nodes in the previous Ï4Y€T.
8 Total weights in neural n€fWOTK. 9 Calculation of error of one iin1age. 10 Four ways to improve the neural netfWOrĂ. 11 Comparison AI vs Machine Learning vs Deep Learning.
12 Imlane Diagram of LSTM Model.-----¿--¿- ¿©5+5++c+cs+css+ 31 Figure 3. 1 The organization of dafaset. 2 Independent variables and dependent variables. 3 The visualization model shows units sold for each product category blusUiii 0112077.
4 The visualization model shows units sold for each product category ACCOPING 1TOTIẨÍHS. 6 SE 9EEk£E*EEkEESEEEEkSEEEEESEEEEkSEEkrkrkrkrkrkrkrkrkrvrk 42 Figure 3. 5 The visualization model shows units sold for each product category according Weelkay. Họ HH go 44 Figure 3.
6 The visualization model shows units sold for each product category EU) SEN. 7 The visualization model shows units sold for each product category ACCOPING SỈOTC. 8 The visualization model shows units sold for each product category ACCOFCING SA. (St TH TH TH HT HT HT ngư 47 Figure 3.
9 Actual vs Predicted on Validation Data— ARIMA Model (1800-1950). 10 Actual vs Predicted on Validation Data— ARIMA Model (810-846). 11 Actual vs Predicted on Validation Data— ARIMA Model (991-999). 12 Actual vs Predicted on Validation Data— LSTM Model (5 years).
13 Actual vs Predicted on Validation Data— LSTM Model (7/2014-3/2016)50 Figure 3. 15 Forecasting model in the future — LSTM model.-‹--‹---- 51 LIST OF ABBREVIATIONS Order | Abbreviation Description 1 EDA Exploratory Data Analysis 2 LSTM Long Short-Term Memory 3 ARIMA Autoregressive Integrated Moving Average 4 ANOVA Analysis of Variance 5 RMSE Root-Mean-Square Error 6 MAE Mean Absolute Error 7 SARIMA Seasonal Autoregressive Integrated Moving Average Abstract The development of life leads to an increase in the demand for people's living, so the retail market grows strongly with big rulers in the retail industry like Amazon, Alibaba, Walmart, eBay, and others all over the world. For any offline or online firm, financial data analysis of sales is essential. It aids in forecasting future values and aids in understanding sales, earnings, and losses.
Therefore, our thesis proposes a solution new method applying advanced techniques to analyze and forecast the unit sales in the unit sales of Walmart retail goods dataset to avoid the reduced profits aspect. By examining the sales data of Walmart stores that are geographically located in various locations and doing a sales projection for the upcoming 28 days, we aim to identify the impact factors on the retail store business in this study. Every day, there are untold numbers of product and income transactions at Walmart locations. The most crucial choice for managers is maintaining the equilibrium between inventory and client demand due to their quick transaction rates.
The retail networks are supported by sales forecasts, allowing for effective resource management. In addition, we presented the proposed approach using Exploratory Data Analysis (EDA) and ANOVA testing statistical method to analyze influence factors and build a predicted model by machine learning (ARIMA-Autoregressive integrated moving average) model and deep learning (LSTM- Long short-term memory) model. Overall, our results show that the LSTM model achieves the lowest error metrics than the ARIMA model of the tested models. With an (RMSE-Root mean square error) of 0.643, LSTM succeeds in outperforming than the RMSE of ARIMA 0.
The LSTM model was applied across the board to anticipate sales units over the next 28 days. The Walmart can make decisions based on this forecast to support the economy going forward and prevent losses. Keywords: EDA, ANOVA, ARIMA, LSTM, RMSE, machine learning, deep learning Problem statement Accurate sales forecasting is essential to any successful retail firm in the cutthroat business environment of today, where sales are of utmost importance to businesses. Future sales are impacted by a number of important things.
These elements can be found by looking at the sales trends of a retail store's overall sales or the sales of a particular product.