微信客服：xiaoxionga100
微信客服：ITCS521
R代写-DPBS1190
时间：2022-04-01
Assessment 4: Group Project Report (30%) ---- Group 1
1
UNSW Diploma
DPBS1190
DATA, INSIGHTS AND
DECISIONS
Assessment 4:
Group Project Report (30%)
Lecturer: Humayun Murshed
Group 1 Members:
Jingyuan Xie (Maggie) z5286680
Zhouyang Jin (Linda) z5286248
Yukang Shao (Zack) z5286242
Zheng Li (Eason) z5337063
Assessment 4: Group Project Report (30%) ---- Group 1
2
DPBS1190 Team 1 Contract:
July 16, 2021
Overview
We, the members of group 1 agree to the following plan of action regarding our work toward the
group assignment tasks.
1. Group Members
zID Name Date
Z5286680 Jingyuan Xie 7/16/2021
Z5286242 Yukang Shao 7/16/2021
Z5286248 Zhouyang Jin 7/16/2021
Z5337063 Zheng Li 7/16/2021
2. Allocation of tasks
In general
Tasks
Group Member
Executive summary (150) -- End of Week 12 Linda
Introduction (200) -- End of Week 12 Maggie
Project goals (100) -- End of Week 12 Eason
Methods (150) -- End of Week 12 Eason
Major body (1000) -- August 2nd All members
Recommended actions (300) -- End of Week 12 Zack
Reflection (100) -- End of Week 12 Linda
Assessment 4: Group Project Report (30%) ---- Group 1
3
For main body
Main body
Group Member
Line chart and analysis Linda
Bubble plot and analysis Maggie
Linear regression and analysis Eason, Maggie, Zack
Forecasting model and analysis Linda, Eason, Zack
Online research All members
3. Meetings & Communication
Number of weekly online meetings 3 meeting for 1h
Person coordinating the meeting for each. All members
Who will summarize decisions, when will he/she make them available
to all members?
Maggie, at the end of each
meeting to WeChat group
4. Work & Deadlines
How will the group come to agreement on a topic? Zoom Meeting or WeChat
When will you make a final decision on a topic? Week 10
Who will collate the draft submissions and then circulate it for
the group to comment on?
Eason
Who will prepare and submit the final submission in Turnitin? Maggie
5. Confirming the original work of group member
We, group 1 of DPBS1190 T11A confirm all the work by our group members are original.
Assessment 4: Group Project Report (30%) ---- Group 1
4
zID Name Signature Date
Z5286680 Jingyuan Xie
7/16/2021
Z5286242 Yukang Shao
7/16/2021
Z5286248 Zhouyang Jin
7/16/2021
Z5337063 Zheng Li
7/16/2021
6. Penalties
What happens if members don’t meet agreed-to deadlines?
If some group members are going to miss the agreed-to deadline; other group members should ask
them whether they are facing difficulties to keep them on the right track. If some group members
deliberately do not finish the task before deadlines, the group can decide to write a written
evaluation of the member's work and pass it into the instructor with the paper.
What happens if members do not contribute / come to meetings?
All the group members should comply with the third article about meetings & communication in
our group contract. The person who is responsible for coordinating meetings needs to schedule
them three days earlier. Therefore, we allow all the group members to plan their schedules and
attend the meeting on time. Schedules might change depending on the emergency incident.
However, if one group member is absent from group meetings over two times without earlier
notice, we would not write their name on the project (check this one with the tutor) and report on
our course convener for further consideration.
Assessment 4: Group Project Report (30%) ---- Group 1
5
Abstract
This report provides predictive analysis models based on sales data for superstore Peoples Pride
under the COVID19 situation. The purpose of this report is to recommend an action plan and
understand the impact on the whole retail market by analyzing the data and searching for
information online. The line charts were used to interpret the trend between confirmed and tested
cases of COVID. In addition, utilizing bubble plots shows the relationship between daily tests,
confirmed cases, and the sale of essential goods. Intuitive assumptions are made to build linear
regression and forecasting models with more actionable insights. Last but not least, to meet our
project’s purpose, the recommendation is provided to improve the current retail situation.
Introduction
There are five dimensions to consider for the role of big data in retailing (Eric et al., 2017). For the
customers' dimension, it is reported that firms can track customers and link transactions over time
(Dekimpe & Hanssens 2000). Other than that, according to Eric et al. (2017), firms can have more
columns about each row to link customer transaction data, e.g., CRM system. For the products
side, product information in marketing can be defined by attributes and levels. The business can
analyze downstream analysis using rows and columns dimensions of product information data
(Ailawadi et al., 2013).
From a times perspective, Eric (2017) explores the benefits of the third add-in dimension:
real-time data makes up for the limitation of historical data, including approximation data, making
decisions more accurate regarding customers behavior and accessible to POS systems and CRM
databases. From the location, nowadays, the ability to use the spatial location of the customer in a
dataset at any given point in time has opened a whole new avenue for retailers to change what
offer to make enabling hyper-targeting (Larson, Bradlow, and Fader 2005). For the channel part,
the study shows that the collection, integration, and analysis of omnichannel data is preferred Eric
et al. (2017).
The potential challenge or notice found by Eric et al. (2017) is that most old data does not reflect
customer needs anymore. The partial dataset can be misleading to the retailers as well.
Project Goals
The report focuses on analyzing the impact of the COVID-19 outbreak on the global retail market.
Through online research, the report learns whether the government’s choice of lockdown in the
rising infection rate causes panic buying and its impact on the sales of life and medical necessities.
Then analyze the similarities and differences with the information provided by the predictive
analysis and find the trend of variables in the future. Provide practical advice for the retail market
in order to quickly recover the economy.
The report carries out visual analysis and predictive analysis of data through R. The report
analyzes the prevalence of COVID-19 using a line chart to analyze the daily number of tests and
Assessment 4: Group Project Report (30%) ---- Group 1
6
the daily number of confirmed cases over time. A bubble plot is used to show three numerical
variables, which show the effect of infection rate on the sales of essential goods. In addition,
simple linear regression models are used to analyze the relationship between specific variables and
infection rates. The quantitative forecasting model is based on historical data of the time series.
We will focus on four quantitative forecasting models to forecast for the four months using the
forecasting model to forecast three critical variables.
R, which supports the data analysis in this report, is free and open-source software. R is explicitly
designed for statistics and data analysis, with high compatibility with the computer operating
system. The logic of R is easy to understand, suitable for professional and non-professional people
to use and interpret. Therefore, R is chosen as the tool for this report.
Analyzation
1. Online Research
Since the outbreak of COVID-19, the retail market has experienced significant adverse impacts.
All kinds of consumer spending are being squeezed. The sales volume of life necessities
experiences a fluctuating trend; they increase positively at the beginning and decrease
dramatically during the latter period, according to the Our World in Data (2021). Under this
situation, however, the Pharmaceutical related retail market breaks the trend (Evans, Santos &
Ford, 2020). The report will analyze the relevant data further to illustrate the pandemic's impact on
the retail market.
2. Line Chart
(Chart 1: Line Chart for Test and Confirmed)
Assessment 4: Group Project Report (30%) ---- Group 1
7
The line chart above displays the daily test cases (blue) and confirmed cases (red) of COVID-19
over a particular period (2020.1.30-2021.6.30). In general, both lines tend to move in the same
direction. However, the confirmed cases show a short-term fluctuation contrasted with the testing
number around 14th December 2020. It may be due to COVID-19 out-of-control situation. The
overall peak point is falling and shows a downward trend after the peak at the end of October 2020.
One reason is that the temperature change with the season may reduce the transmission rate of
infection, thus reducing the infection rate of individuals. In addition, it can also be due to the
mature system of community lockdown policy.
3. Bubble Plot
(Chart 2: Bubble plot for sales of essential goods)
The bubble plot shows a positive linear relationship between test cases and confirmed cases. In
general, a lower confirmed case number and lower test number have higher sales of essential
goods, except for the test numbers at 30000.
In the beginning, the size of the bubble plot is large, which means that the sales of purchasing
essential goods are large. Later, as confirmed cases and test cases gradually increase, the sales of
essential goods turn into a declining trend.
It is possible that in January 2020, as the outbreak began to be confirmed, residents realized the
seriousness of the situation and began to stock up and buy essential goods. The published
lockdown policies may be in place and made residents traveling less, resulting in fewer sales of
essential goods. The mid-term exception maybe because the consumption of essential goods or the
epidemic situation has improved. The lockdown policy has eased, so people go out to buy goods.
Assessment 4: Group Project Report (30%) ---- Group 1
8
4. Regression
Assumption
The report uses infection rate as a benchmark and simple linear regression to find the relationship
between an individual variable and the benchmark. To ensure the accuracy of the model, we set all
residuals to 0. A significant level of 0.05 is used to test whether the infection rate is statistically
significant. Meanwhile, we selected the median infection rate of 15574 for prediction since the
median is not affected by extreme values.
Medicine
According to the database, the model equation between medicine and the confirmed case is set to:
= + ∗ +
(Table 1: Regression model for Sales of Medicine and Confirmed Case)
The result is summarized as:
= 1313 + 0.001437 ∗ +
The coefficients show that the daily sales of medicine are $1313 when the number of confirmed
cases is 0. A slope of 0.001437 means that for every additional confirmed case, sales of medicine
increase by $0.001437. Based on a median of 15574 confirmed cases between January 2020 and
June 2021, the daily sales would be $1335, which increases 102%. It is in line with the massive
increase in medicine sales cited in online research. The adjusted R-squared is 0.025 means the
model is a poor fit. The p-value of 0.00018 is much smaller than the significant level of 0.05. The
infection rate is statistically significant.
Totalsales
(Table 2: Regression model for Total Sales and Confirmed Case)
The result is summarized as:
= 72229 − 1.6246 ∗ +
Assessment 4: Group Project Report (30%) ---- Group 1
9
The slope shows that as confirmed cases increase by $1, total sales are predicted to decrease by
$1.6246. Based on a median of 15574 confirmed cases, the daily sales would be $46928,
decreasing 97.64% compared to the first day sales. It can be related to the depressing situation in
retail marketing during COVID-19 cited in online research (Evans, Santos & Ford, 2020). The
adjusted R-squared is 0.0064 means the model is a poor fit. It may be due to the epidemic under
control, and other factors may affect total sales.
Essentialgoods
(Table 3: Regression model for Sales of Essential goods and Confirmed Case)
The model equation is:
= 71484 − 1.6217 ∗ +
The p-value of essential goods is 0.0367, which is less than 0.05 means it is statistically significant.
The coefficient relationship between confirmed and essential goods is that increasing the unit of
essential goods would decrease 1.6217 units of confirmed, so there is a negative relationship. It is
noteworthy that the adjusted R-square is 0.006511, which means the model is inferior.
5. Forecast
Facemask
(Table 4: Forecast model for Facemask)
Assessment 4: Group Project Report (30%) ---- Group 1
10
The information for an estimate of deals in facemask in the accompanying four months shows that
the pattern is increment the 128.08 each month. It is negative worth between seasons 4 and 10,
which shows the adjustment of deals for a facemask with the change for the season.
(Chart 3: Time Series for Facemask)
(Table 5: Forecast Model for Facemask over Next 4 Months)
Additionally, it will ascend around 150 consistently from July to October. The retail market
available to be purchased facemask will affect via abnormality since residents do not accept
facemasks in the specific season. According to the prediction model made from the dataset, the
sales volume of masks did not wholly match the online confidence.
Handsanitizer
(Table 6: Forecast Model for Handsanitizer)
Assessment 4: Group Project Report (30%) ---- Group 1
11
According to the coefficient trend, hand sanitizer sales are expected to increase by 223.523 per
month. Season 2 has the highest sales of hand sanitizer based on the season section because the
expectation of season 2 exceeds 6.16 compared with the benchmark of season 1.
Besides, season 9 faces the lowest sales of hand sanitizer since its anticipation is lower 364.49
than the benchmark of season 1. One reason causes this situation might be that the popularity of
vaccines increases dramatically. It is noteworthy that the expectation of season 2 is not significant.
(Chart 4: Time Series for Handsanitizer)
(Table 7: Forecast Model for Handsanitizer over Next 4 Months)
The time series diagram for Handsanitizer demonstrates positive sales value over the next four
months due to hand sanitizers are effective in preventing viruses. There is 95% confidence that the
actual value will occur within the forecast range.
Assessment 4: Group Project Report (30%) ---- Group 1
12
Totalsales
(Table 8: Forecast Model for Total Sales)
The table shows that seasonality has no significant impact on the total sales. The trend of total
sales is expected to decline by $67702 per month. In season 3, the total sales increased by $28739,
which means people were able to hoard goods during this period.
(Chart 5: Time Series for Total Sales)
(Table 9: Forecast Model for Total Sales over Next 4 Months)
Over the next four months, total sales will continue to decline, but the situation will gradually
improve. Actual value fall within the forecast range with 95% confidence.
Assessment 4: Group Project Report (30%) ---- Group 1
13
Limitations
There are some existing limitations of the dataset in the process of analysis. The variables of the
dataset don do not show a complete picture of the whole covid-19 situation. The database only
measures the specific. The measurement range of the observed value, such as the type of medicine,
should be noted as well.
In addition, when subject attrition is efficient, patterns might mirror the changing idea of the
example, maybe the elements of the marvels being scrutinized (Itzhak and Arthur,2007). Other
bewildering impacts might be made by verifiable occasions and changes in measures or recording
practices of factors over the long haul (Itzhak and Arthur,2007). For example, the dataset shows
that these data can reflect in one period.
Recommendations
One of the most efficient methods to solve this problem is to advocate the public to implement
epidemic prevention strategies such as increasing the publicity effort of epidemic items, discount
for epidemic items, providing online delivery, and focusing on employee infection. The first
strategy is that the retail market needs publicity to increase total sales during the epidemic. The
forecast model of facemask demonstrates that the number of facemasks sold is down from seasons
4 to 10. One reason is maybe citizens lack protection consciousness. Another reason is that people
intend to stay at home instead of buying goods outside to prevent infection.
Nevertheless, there is no good way in the long term, and the face mask can provide freedom to the
public to go out. Therefore, it is suggested that the retail market needs advertisements to remind
individuals about the importance of facemasks. The advertisements format can be including
pictures, videos, and messages.
Next, according to the regression between essential goods and confirmed, the sales of essential
goods would decrease due to increasing the number of confirmed cases. A reason might be that
people would not intend to go out frequently due to COVID-19 because most people choose to
purchase goods when they do not have enough stock. However, retail marketing should provide
the discount to customers at this period, for the public would intend to increase their store because
of discount.
Finally, the diagram above shows that the total sale decreases with the increase of confirmed cases.
Therefore, it is recommended that the retail market give more opportunities to transfer some of the
physical offline business patterns into online sales. The retail market can apply to the government
for delivery, which prevents businesses from having no or low income during the lockdown and
satisfies people’s demand for consumption. Developing a new clicks-and-mortar business model
that combines online e-commerce with physical retail outlets may solve the inconvenience of
residents not getting access to physical stores. One thing that needs to be noted is that businesses
need to take preventive measures to avoid a second epidemic outbreak because they must contact
many customers. Specifically, it is required to wear a facemask and show the health QR code
Assessment 4: Group Project Report (30%) ---- Group 1
14
before getting into the store. All the goods and productions need to be disinfected daily by the
staff.
Reflection
In the process of analysis, we took in the fundamental examination and factual apparatuses for
getting clients, markets, contenders, and ourselves. Apply accurately and information examination
abilities to genuine enormous informational indexes, zeroing in on applications instead of
techniques. We figured out how to depict, gather, and examine business information from various
fields like money, advertising, the executives, and hazard examination. Information experiences
and ideas would then be viably imparted to non-specialized crowds, such as data visualization. All
through the abilities above, it can help we are applying to genuine cases. Like this project, we are
utilizing various models to demonstrate our presumptions and give conclusions.
Appendix
Chart 1: Line Chart for Test and Confirmed
library(readxl)
MyData <- read_excel("C:/Users/Linda/Desktop/MyData.xlsx")
View(MyData)
library(ggplot2)
ggplot()+
geom_line(data=MyData,mapping=aes(x=Date,y=Test),color="blue")+
geom_line(data=MyData,mapping=aes(x=Date,y=Confirmed),color="red")
+
labs(title="Linechart for Test and Confirmed",y = "Daily number of
Test and confirmed cases")
Chart 2: Bubble Plot for the Sales of Essentialgoods
library(readxl)
myData <- read_excel("C:/Users/admin/Desktop/1190 Group Work/myData.x
lsx")
View(myData)
plot(myData$Confirmed~myData$Test)
symbols(myData$Confirmed~myData$Test,circles = myData$Essentialgoods,
inches = 0.4,bg="red",main = "A Bubble plot for sales of essential go
ods",xlab="Test", ylab="Confirmed" )
abline(lm(myData$Confirmed~myData$Test),col='black')
Table 1: Regression for Medicine and Confirmed
library(readxl)
myData <- read_excel("C:/Users/admin/Desktop/1190 Group Work/myData.x
Assessment 4: Group Project Report (30%) ---- Group 1
15
lsx")
View(myData)
Model=lm(Medicine~Confirmed,data=myData)
summary(Model)
##
## Call:
## lm(formula = Medicine ~ Confirmed, data = myData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -374.08 -51.80 1.25 57.14 118.77
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.313e+03 7.163e+00 183.235 < 2e-16 ***
## Confirmed 1.437e-03 3.808e-04 3.773 0.00018 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 74.08 on 516 degrees of freedom
## Multiple R-squared: 0.02685, Adjusted R-squared: 0.02496
## F-statistic: 14.23 on 1 and 516 DF, p-value: 0.0001801
summary(myData$Confirmed)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1033 10722 15574 16756 21731 39542
predict(Model,data.frame(Confirmed=15574))
## 1
## 1334.977
Table 2: Regression for Totalsales and Confirmed
library(readxl)
myData <- read_excel("C:/Users/jingy/OneDrive/words/myData.xlsx")
View(myData)
Model=lm(Totalsales~Confirmed,data=myData)
summary(Model)
##
## Call:
## lm(formula = Totalsales ~ Confirmed, data = myData)
##
## Residuals:
Assessment 4: Group Project Report (30%) ---- Group 1
16
## Min 1Q Median 3Q Max
## -64943 -42352 -29480 -11085 1918457
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 72229.4047 14613.3325 4.943 1.04e-06 ***
## Confirmed -1.6246 0.7769 -2.091 0.037 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 151100 on 516 degrees of freedom
## Multiple R-squared: 0.008403, Adjusted R-squared: 0.006482
## F-statistic: 4.373 on 1 and 516 DF, p-value: 0.037
summary(myData$Confirmed)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1033 10722 15574 16756 21731 39542
predict(Model,data.frame(Confirmed=15574))
## 1
## 46928.1
Table 3: Regression for Essentialgoods and Confirmed
ibrary(readxl)
myData <- read_excel("C:/Users/Orb/Desktop/myData.xlsx")
View(myData)
packages.required <- c("data.table", "ggplot2","magrittr","dplyr")
new.packages <- packages.required[!(packages.required %in% installed.
packages()[,"Package"])]
library(readxl)
myData <- read_excel("C:/Users/Orb/Desktop/myData.xlsx")
View(myData)
Model=lm(Essentialgoods~Confirmed,data=myData)
summary(Model)
##
## Call:
## lm(formula = Essentialgoods ~ Confirmed, data = myData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -64471 -42048 -29401 -11151 1914684
##
Assessment 4: Group Project Report (30%) ---- Group 1
17
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 71483.7186 14561.5344 4.909 1.23e-06 ***
## Confirmed -1.6217 0.7741 -2.095 0.0367 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 150600 on 516 degrees of freedom
## Multiple R-squared: 0.008433, Adjusted R-squared: 0.006511
## F-statistic: 4.388 on 1 and 516 DF, p-value: 0.03667
summary(myData$Confirmed)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1033 10722 15574 16756 21731 39542
Table 4: Forecast for Facemasks
library(readxl)
MyData <- read_excel("C:/Users/Linda/Desktop/MyData.xlsx")
View(MyData)
library(forecast)
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
head(MyData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
## <
dbl>
## 1 2020-01-30 00:00:00 1260 1033. 106 1989008. 198
4492.
## 2 2020-01-31 00:00:00 1386 1137. 81 749637. 74
6415
## 3 2020-02-01 00:00:00 1525. 1250. 60 1600316. 159
2408.
## 4 2020-02-02 00:00:00 1677. 1375. 101 171372. 16
6895.
## 5 2020-02-03 00:00:00 1845. 1513. 49 6079.
5700.
## 6 2020-02-04 00:00:00 2029. 1664. 51 1000644. 99
8389.
Assessment 4: Group Project Report (30%) ---- Group 1
18
## # ... with 6 more variables: Handsanitizer, Facemask ,
## # Toilettissue, Medicine , Lockdown , Panicbuyin
g
tail(MyData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
## <
dbl>
## 1 2021-06-25 00:00:00 8833. 7420. 34 4012.
3518.
## 2 2021-06-26 00:00:00 8568. 7197. 36 25194. 2
4746.
## 3 2021-06-27 00:00:00 8311. 6981. 9 57637. 5
7031.
## 4 2021-06-28 00:00:00 8062. 6772. 31 11709. 1
0726.
## 5 2021-06-29 00:00:00 7820. 6569. 14 12821. 1
2459.
## 6 2021-06-30 00:00:00 7585. 6372. 21 58847. 5
8082.
## # ... with 6 more variables: Handsanitizer, Facemask ,
## # Toilettissue, Medicine , Lockdown , Panicbuyin
g
newData=ts(MyData$Facemask,start=c(2020,1,30),end=c(2021,6,30),frequ
ency=12)
TSReg=tslm(newData~trend+season)
summary(TSReg)
##
## Call:
## tslm(formula = newData ~ trend + season)
##
## Residuals:
## [1] -1.690e+02 -2.159e+02 -6.213e+02 1.131e+03 -4.210e+01 -8.263
e+01
## [7] -6.395e-14 -6.395e-14 -1.421e-14 -4.263e-14 -1.208e-13 6.395
e-14
## [13] 1.690e+02 2.159e+02 6.213e+02 -1.131e+03 4.210e+01 8.263
e+01
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
Assessment 4: Group Project Report (30%) ---- Group 1
19
## (Intercept) 2396.95 655.06 3.659 0.0146 *
## trend 128.08 40.24 3.183 0.0245 *
## season2 36.60 837.34 0.044 0.9668
## season3 437.56 840.23 0.521 0.6248
## season4 -1312.86 845.04 -1.554 0.1810
## season5 -131.52 851.72 -0.154 0.8833
## season6 -75.89 860.23 -0.088 0.9331
## season7 -136.25 1024.34 -0.133 0.8994
## season8 -106.47 1025.13 -0.104 0.9213
## season9 -68.79 1027.50 -0.067 0.9492
## season10 -22.83 1031.43 -0.022 0.9832
## season11 31.83 1036.91 0.031 0.9767
## season12 95.64 1043.91 0.092 0.9306
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 836.4 on 5 degrees of freedom
## Multiple R-squared: 0.7606, Adjusted R-squared: 0.186
## F-statistic: 1.324 on 12 and 5 DF, p-value: 0.4018
forecast(TSReg,h=4)
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jul 2021 4694.232 2808.674 6579.791 1410.111 7978.354
## Aug 2021 4852.095 2966.537 6737.654 1567.974 8136.217
## Sep 2021 5017.852 3132.293 6903.410 1733.730 8301.973
## Oct 2021 5191.896 3306.338 7077.455 1907.775 8476.018
plot(forecast(TSReg,h=4),xlab='Date',ylab='Facemask',main='Time Seri
es for Facemask')
Table 5: Forecast for Handsanitizer
library(readxl)
myData <- read_excel("C:/Users/Orb/Desktop/myData.xlsx")
View(myData)
head(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
## <
dbl>
## 1 2020-01-30 00:00:00 1260 1033. 106 1989008. 198
4492.
## 2 2020-01-31 00:00:00 1386 1137. 81 749637. 74
Assessment 4: Group Project Report (30%) ---- Group 1
20
6415
## 3 2020-02-01 00:00:00 1525. 1250. 60 1600316. 159
2408.
## 4 2020-02-02 00:00:00 1677. 1375. 101 171372. 16
6895.
## 5 2020-02-03 00:00:00 1845. 1513. 49 6079.
5700.
## 6 2020-02-04 00:00:00 2029. 1664. 51 1000644. 99
8389.
## # ... with 6 more variables: Handsanitizer, Facemask ,
## # Toilettissue, Medicine , Lockdown , Panicbuyin
g
tail(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
## <
dbl>
## 1 2021-06-25 00:00:00 8833. 7420. 34 4012.
3518.
## 2 2021-06-26 00:00:00 8568. 7197. 36 25194. 2
4746.
## 3 2021-06-27 00:00:00 8311. 6981. 9 57637. 5
7031.
## 4 2021-06-28 00:00:00 8062. 6772. 31 11709. 1
0726.
## 5 2021-06-29 00:00:00 7820. 6569. 14 12821. 1
2459.
## 6 2021-06-30 00:00:00 7585. 6372. 21 58847. 5
8082.
## # ... with 6 more variables: Handsanitizer, Facemask ,
## # Toilettissue, Medicine , Lockdown , Panicbuyin
g
library(forecast)
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
newData=ts(myData$Handsanitizer,start=c(2020,01,30),end=c(2021,06,3
0),frequency=12)
TSReg=tslm(newData~trend+season)
summary(TSReg)
Assessment 4: Group Project Report (30%) ---- Group 1
21
##
## Call:
## tslm(formula = newData ~ trend + season)
##
## Residuals:
## [1] 1.543e+02 3.563e+01 8.776e-01 -3.256e+01 -6.431e+01 -9.395
e+01
## [7] -1.066e-14 -1.066e-14 1.066e-14 -2.132e-14 1.421e-14 0.000
e+00
## [13] -1.543e+02 -3.563e+01 -8.776e-01 3.256e+01 6.431e+01 9.395
e+01
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 732.166 97.954 7.475 0.000677 ***
## trend 223.523 6.017 37.147 2.66e-07 ***
## season2 6.160 125.211 0.049 0.962669
## season3 -60.513 125.644 -0.482 0.650416
## season4 -116.290 126.363 -0.920 0.399646
## season5 -160.323 127.362 -1.259 0.263676
## season6 -191.687 128.635 -1.490 0.196367
## season7 -330.395 153.175 -2.157 0.083505 .
## season8 -357.275 153.293 -2.331 0.067156 .
## season9 -364.490 153.647 -2.372 0.063774 .
## season10 -350.075 154.235 -2.270 0.072461 .
## season11 -311.866 155.054 -2.011 0.100475
## season12 -247.484 156.102 -1.585 0.173732
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 125.1 on 5 degrees of freedom
## Multiple R-squared: 0.9967, Adjusted R-squared: 0.9889
## F-statistic: 127.7 on 12 and 5 DF, p-value: 2.099e-05
forecast(TSReg,h=4)
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jul 2021 4648.710 4366.753 4930.667 4157.619 5139.801
## Aug 2021 4845.353 4563.396 5127.310 4354.262 5336.444
## Sep 2021 5061.661 4779.704 5343.618 4570.570 5552.752
## Oct 2021 5299.599 5017.642 5581.556 4808.508 5790.690
plot(forecast(TSReg,h=4),xlab='Date',ylab='Handsanitizer',main='Time
Series for Handsanitizer')
Assessment 4: Group Project Report (30%) ---- Group 1
22
Table 6: Forecast for Totalsales
library(readxl)
myData <- read_excel("1190 Group Work/myData.xlsx")
View(myData)
head(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
## <
dbl>
## 1 2020-01-30 00:00:00 1260 1033. 106 1989008. 198
4492.
## 2 2020-01-31 00:00:00 1386 1137. 81 749637. 74
6415
## 3 2020-02-01 00:00:00 1525. 1250. 60 1600316. 159
2408.
## 4 2020-02-02 00:00:00 1677. 1375. 101 171372. 16
6895.
## 5 2020-02-03 00:00:00 1845. 1513. 49 6079.
5700.
## 6 2020-02-04 00:00:00 2029. 1664. 51 1000644. 99
8389.
## # ... with 6 more variables: Handsanitizer, Facemask ,
## # Toilettissue, Medicine , Lockdown , Panicbuyin
g
tail(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
## <
dbl>
## 1 2021-06-25 00:00:00 8833. 7420. 34 4012.
3518.
## 2 2021-06-26 00:00:00 8568. 7197. 36 25194. 2
4746.
## 3 2021-06-27 00:00:00 8311. 6981. 9 57637. 5
7031.
## 4 2021-06-28 00:00:00 8062. 6772. 31 11709. 1
0726.
## 5 2021-06-29 00:00:00 7820. 6569. 14 12821. 1
2459.
## 6 2021-06-30 00:00:00 7585. 6372. 21 58847. 5
Assessment 4: Group Project Report (30%) ---- Group 1
23
8082.
## # ... with 6 more variables: Handsanitizer, Facemask ,
## # Toilettissue, Medicine , Lockdown , Panicbuyin
g
library(forecast)
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
newData=ts(myData$Totalsales,start=c(2020,01,30),end=c(2021,06,30),f
requency=12)
TSReg=tslm(newData~trend+season)
summary(TSReg)
##
## Call:
## tslm(formula = newData ~ trend + season)
##
## Residuals:
## [1] 5.844e+05 -3.881e+04 3.024e+05 -4.707e+05 -4.043e+05 2.704
e+04
## [7] 4.366e-11 1.455e-11 2.910e-11 1.455e-11 7.276e-11 5.821
e-11
## [13] -5.844e+05 3.881e+04 -3.024e+05 4.707e+05 4.043e+05 -2.704
e+04
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1472315 448610 3.282 0.0219 *
## trend -67702 27558 -2.457 0.0575 .
## season2 -548464 573442 -0.956 0.3828
## season3 28739 575425 0.050 0.9621
## season4 -559460 578715 -0.967 0.3781
## season5 -723405 583290 -1.240 0.2699
## season6 -92496 589120 -0.157 0.8814
## season7 -992727 701509 -1.415 0.2162
## season8 -877908 702050 -1.250 0.2665
## season9 -838231 703670 -1.191 0.2870
## season10 -582707 706363 -0.825 0.4470
## season11 -724656 710116 -1.020 0.3543
## season12 -533442 714913 -0.746 0.4891
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
Assessment 4: Group Project Report (30%) ---- Group 1
24
## Residual standard error: 572800 on 5 degrees of freedom
## Multiple R-squared: 0.7223, Adjusted R-squared: 0.05591
## F-statistic: 1.084 on 12 and 5 DF, p-value: 0.5007
forecast(TSReg,h=4)
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jul 2021 -806755.2 -2098058 484547.3 -3055847 1442337
## Aug 2021 -759639.3 -2050942 531663.2 -3008731 1489452
## Sep 2021 -787664.6 -2078967 503637.9 -3036756 1461427
## Oct 2021 -599842.7 -1891145 691459.8 -2848935 1649249
plot(forecast(TSReg,h=4),xlab='Date',ylab='Totalsales',main='Time Se
ries for Totalsales')
Reference
[1] Eric, T, B., Manish,G., Praveen,K., Sudhir, Voleti. (March 2017). The Role of Big Data and
Predictive Analytics in Retailing. Volume 93, Issue 1. Available from
[2] Dekimpe, Marnik G. and Dominique M. Hanssens (2000), “Time-series Models in Marketing:
Past, Present and Future,” International Journal of Research in Marketing, 17 (2), 183–93.
[3] Ailawadi, Kusum L., Donald R. Lehmann and Scott A. Neslin (2003), “Revenue Premium as
an Outcome Measure of Brand Equity,” Journal of Marketing, 67 (4), 1–17.
[4] Larson, Jeffrey S., Eric T. Bradlow and Peter S. Fader (2005), “An Exploratory Look at
Supermarket Shopping Paths,” International Journal of Research in Marketing, 22 (4), 395–414.
[5] Ritchie, H., Ortiz-Ospina, E., Beltekian, D., Mathieu, E., Hasell, J., Macdonald, B., Giattino,
C., Appel, C., Rodés-Guirao, L., & Roser, M. (March 5, 2020). Coronavirus (covid-19)
VACCINATIONS - statistics and research. Our World in Data. Available from
[6] Evans, A., Santos, E., Ford, M. (May 1, 2020). Impact of COVID-19 on Vitamins, Minerals
and Supplements. Volume XXII, issue 32. Available from
1
UNSW Diploma
DPBS1190
DATA, INSIGHTS AND
DECISIONS
Assessment 4:
Group Project Report (30%)
Lecturer: Humayun Murshed
Group 1 Members:
Jingyuan Xie (Maggie) z5286680
Zhouyang Jin (Linda) z5286248
Yukang Shao (Zack) z5286242
Zheng Li (Eason) z5337063
Assessment 4: Group Project Report (30%) ---- Group 1
2
DPBS1190 Team 1 Contract:
July 16, 2021
Overview
We, the members of group 1 agree to the following plan of action regarding our work toward the
group assignment tasks.
1. Group Members
zID Name Date
Z5286680 Jingyuan Xie 7/16/2021
Z5286242 Yukang Shao 7/16/2021
Z5286248 Zhouyang Jin 7/16/2021
Z5337063 Zheng Li 7/16/2021
2. Allocation of tasks
In general
Tasks
Group Member
Executive summary (150) -- End of Week 12 Linda
Introduction (200) -- End of Week 12 Maggie
Project goals (100) -- End of Week 12 Eason
Methods (150) -- End of Week 12 Eason
Major body (1000) -- August 2nd All members
Recommended actions (300) -- End of Week 12 Zack
Reflection (100) -- End of Week 12 Linda
Assessment 4: Group Project Report (30%) ---- Group 1
3
For main body
Main body
Group Member
Line chart and analysis Linda
Bubble plot and analysis Maggie
Linear regression and analysis Eason, Maggie, Zack
Forecasting model and analysis Linda, Eason, Zack
Online research All members
3. Meetings & Communication
Number of weekly online meetings 3 meeting for 1h
Person coordinating the meeting for each. All members
Who will summarize decisions, when will he/she make them available
to all members?
Maggie, at the end of each
meeting to WeChat group
4. Work & Deadlines
How will the group come to agreement on a topic? Zoom Meeting or WeChat
When will you make a final decision on a topic? Week 10
Who will collate the draft submissions and then circulate it for
the group to comment on?
Eason
Who will prepare and submit the final submission in Turnitin? Maggie
5. Confirming the original work of group member
We, group 1 of DPBS1190 T11A confirm all the work by our group members are original.
Assessment 4: Group Project Report (30%) ---- Group 1
4
zID Name Signature Date
Z5286680 Jingyuan Xie
7/16/2021
Z5286242 Yukang Shao
7/16/2021
Z5286248 Zhouyang Jin
7/16/2021
Z5337063 Zheng Li
7/16/2021
6. Penalties
What happens if members don’t meet agreed-to deadlines?
If some group members are going to miss the agreed-to deadline; other group members should ask
them whether they are facing difficulties to keep them on the right track. If some group members
deliberately do not finish the task before deadlines, the group can decide to write a written
evaluation of the member's work and pass it into the instructor with the paper.
What happens if members do not contribute / come to meetings?
All the group members should comply with the third article about meetings & communication in
our group contract. The person who is responsible for coordinating meetings needs to schedule
them three days earlier. Therefore, we allow all the group members to plan their schedules and
attend the meeting on time. Schedules might change depending on the emergency incident.
However, if one group member is absent from group meetings over two times without earlier
notice, we would not write their name on the project (check this one with the tutor) and report on
our course convener for further consideration.
Assessment 4: Group Project Report (30%) ---- Group 1
5
Abstract
This report provides predictive analysis models based on sales data for superstore Peoples Pride
under the COVID19 situation. The purpose of this report is to recommend an action plan and
understand the impact on the whole retail market by analyzing the data and searching for
information online. The line charts were used to interpret the trend between confirmed and tested
cases of COVID. In addition, utilizing bubble plots shows the relationship between daily tests,
confirmed cases, and the sale of essential goods. Intuitive assumptions are made to build linear
regression and forecasting models with more actionable insights. Last but not least, to meet our
project’s purpose, the recommendation is provided to improve the current retail situation.
Introduction
There are five dimensions to consider for the role of big data in retailing (Eric et al., 2017). For the
customers' dimension, it is reported that firms can track customers and link transactions over time
(Dekimpe & Hanssens 2000). Other than that, according to Eric et al. (2017), firms can have more
columns about each row to link customer transaction data, e.g., CRM system. For the products
side, product information in marketing can be defined by attributes and levels. The business can
analyze downstream analysis using rows and columns dimensions of product information data
(Ailawadi et al., 2013).
From a times perspective, Eric (2017) explores the benefits of the third add-in dimension:
real-time data makes up for the limitation of historical data, including approximation data, making
decisions more accurate regarding customers behavior and accessible to POS systems and CRM
databases. From the location, nowadays, the ability to use the spatial location of the customer in a
dataset at any given point in time has opened a whole new avenue for retailers to change what
offer to make enabling hyper-targeting (Larson, Bradlow, and Fader 2005). For the channel part,
the study shows that the collection, integration, and analysis of omnichannel data is preferred Eric
et al. (2017).
The potential challenge or notice found by Eric et al. (2017) is that most old data does not reflect
customer needs anymore. The partial dataset can be misleading to the retailers as well.
Project Goals
The report focuses on analyzing the impact of the COVID-19 outbreak on the global retail market.
Through online research, the report learns whether the government’s choice of lockdown in the
rising infection rate causes panic buying and its impact on the sales of life and medical necessities.
Then analyze the similarities and differences with the information provided by the predictive
analysis and find the trend of variables in the future. Provide practical advice for the retail market
in order to quickly recover the economy.
The report carries out visual analysis and predictive analysis of data through R. The report
analyzes the prevalence of COVID-19 using a line chart to analyze the daily number of tests and
Assessment 4: Group Project Report (30%) ---- Group 1
6
the daily number of confirmed cases over time. A bubble plot is used to show three numerical
variables, which show the effect of infection rate on the sales of essential goods. In addition,
simple linear regression models are used to analyze the relationship between specific variables and
infection rates. The quantitative forecasting model is based on historical data of the time series.
We will focus on four quantitative forecasting models to forecast for the four months using the
forecasting model to forecast three critical variables.
R, which supports the data analysis in this report, is free and open-source software. R is explicitly
designed for statistics and data analysis, with high compatibility with the computer operating
system. The logic of R is easy to understand, suitable for professional and non-professional people
to use and interpret. Therefore, R is chosen as the tool for this report.
Analyzation
1. Online Research
Since the outbreak of COVID-19, the retail market has experienced significant adverse impacts.
All kinds of consumer spending are being squeezed. The sales volume of life necessities
experiences a fluctuating trend; they increase positively at the beginning and decrease
dramatically during the latter period, according to the Our World in Data (2021). Under this
situation, however, the Pharmaceutical related retail market breaks the trend (Evans, Santos &
Ford, 2020). The report will analyze the relevant data further to illustrate the pandemic's impact on
the retail market.
2. Line Chart
(Chart 1: Line Chart for Test and Confirmed)
Assessment 4: Group Project Report (30%) ---- Group 1
7
The line chart above displays the daily test cases (blue) and confirmed cases (red) of COVID-19
over a particular period (2020.1.30-2021.6.30). In general, both lines tend to move in the same
direction. However, the confirmed cases show a short-term fluctuation contrasted with the testing
number around 14th December 2020. It may be due to COVID-19 out-of-control situation. The
overall peak point is falling and shows a downward trend after the peak at the end of October 2020.
One reason is that the temperature change with the season may reduce the transmission rate of
infection, thus reducing the infection rate of individuals. In addition, it can also be due to the
mature system of community lockdown policy.
3. Bubble Plot
(Chart 2: Bubble plot for sales of essential goods)
The bubble plot shows a positive linear relationship between test cases and confirmed cases. In
general, a lower confirmed case number and lower test number have higher sales of essential
goods, except for the test numbers at 30000.
In the beginning, the size of the bubble plot is large, which means that the sales of purchasing
essential goods are large. Later, as confirmed cases and test cases gradually increase, the sales of
essential goods turn into a declining trend.
It is possible that in January 2020, as the outbreak began to be confirmed, residents realized the
seriousness of the situation and began to stock up and buy essential goods. The published
lockdown policies may be in place and made residents traveling less, resulting in fewer sales of
essential goods. The mid-term exception maybe because the consumption of essential goods or the
epidemic situation has improved. The lockdown policy has eased, so people go out to buy goods.
Assessment 4: Group Project Report (30%) ---- Group 1
8
4. Regression
Assumption
The report uses infection rate as a benchmark and simple linear regression to find the relationship
between an individual variable and the benchmark. To ensure the accuracy of the model, we set all
residuals to 0. A significant level of 0.05 is used to test whether the infection rate is statistically
significant. Meanwhile, we selected the median infection rate of 15574 for prediction since the
median is not affected by extreme values.
Medicine
According to the database, the model equation between medicine and the confirmed case is set to:
= + ∗ +
(Table 1: Regression model for Sales of Medicine and Confirmed Case)
The result is summarized as:
= 1313 + 0.001437 ∗ +
The coefficients show that the daily sales of medicine are $1313 when the number of confirmed
cases is 0. A slope of 0.001437 means that for every additional confirmed case, sales of medicine
increase by $0.001437. Based on a median of 15574 confirmed cases between January 2020 and
June 2021, the daily sales would be $1335, which increases 102%. It is in line with the massive
increase in medicine sales cited in online research. The adjusted R-squared is 0.025 means the
model is a poor fit. The p-value of 0.00018 is much smaller than the significant level of 0.05. The
infection rate is statistically significant.
Totalsales
(Table 2: Regression model for Total Sales and Confirmed Case)
The result is summarized as:
= 72229 − 1.6246 ∗ +
Assessment 4: Group Project Report (30%) ---- Group 1
9
The slope shows that as confirmed cases increase by $1, total sales are predicted to decrease by
$1.6246. Based on a median of 15574 confirmed cases, the daily sales would be $46928,
decreasing 97.64% compared to the first day sales. It can be related to the depressing situation in
retail marketing during COVID-19 cited in online research (Evans, Santos & Ford, 2020). The
adjusted R-squared is 0.0064 means the model is a poor fit. It may be due to the epidemic under
control, and other factors may affect total sales.
Essentialgoods
(Table 3: Regression model for Sales of Essential goods and Confirmed Case)
The model equation is:
= 71484 − 1.6217 ∗ +
The p-value of essential goods is 0.0367, which is less than 0.05 means it is statistically significant.
The coefficient relationship between confirmed and essential goods is that increasing the unit of
essential goods would decrease 1.6217 units of confirmed, so there is a negative relationship. It is
noteworthy that the adjusted R-square is 0.006511, which means the model is inferior.
5. Forecast
Facemask
(Table 4: Forecast model for Facemask)
Assessment 4: Group Project Report (30%) ---- Group 1
10
The information for an estimate of deals in facemask in the accompanying four months shows that
the pattern is increment the 128.08 each month. It is negative worth between seasons 4 and 10,
which shows the adjustment of deals for a facemask with the change for the season.
(Chart 3: Time Series for Facemask)
(Table 5: Forecast Model for Facemask over Next 4 Months)
Additionally, it will ascend around 150 consistently from July to October. The retail market
available to be purchased facemask will affect via abnormality since residents do not accept
facemasks in the specific season. According to the prediction model made from the dataset, the
sales volume of masks did not wholly match the online confidence.
Handsanitizer
(Table 6: Forecast Model for Handsanitizer)
Assessment 4: Group Project Report (30%) ---- Group 1
11
According to the coefficient trend, hand sanitizer sales are expected to increase by 223.523 per
month. Season 2 has the highest sales of hand sanitizer based on the season section because the
expectation of season 2 exceeds 6.16 compared with the benchmark of season 1.
Besides, season 9 faces the lowest sales of hand sanitizer since its anticipation is lower 364.49
than the benchmark of season 1. One reason causes this situation might be that the popularity of
vaccines increases dramatically. It is noteworthy that the expectation of season 2 is not significant.
(Chart 4: Time Series for Handsanitizer)
(Table 7: Forecast Model for Handsanitizer over Next 4 Months)
The time series diagram for Handsanitizer demonstrates positive sales value over the next four
months due to hand sanitizers are effective in preventing viruses. There is 95% confidence that the
actual value will occur within the forecast range.
Assessment 4: Group Project Report (30%) ---- Group 1
12
Totalsales
(Table 8: Forecast Model for Total Sales)
The table shows that seasonality has no significant impact on the total sales. The trend of total
sales is expected to decline by $67702 per month. In season 3, the total sales increased by $28739,
which means people were able to hoard goods during this period.
(Chart 5: Time Series for Total Sales)
(Table 9: Forecast Model for Total Sales over Next 4 Months)
Over the next four months, total sales will continue to decline, but the situation will gradually
improve. Actual value fall within the forecast range with 95% confidence.
Assessment 4: Group Project Report (30%) ---- Group 1
13
Limitations
There are some existing limitations of the dataset in the process of analysis. The variables of the
dataset don do not show a complete picture of the whole covid-19 situation. The database only
measures the specific. The measurement range of the observed value, such as the type of medicine,
should be noted as well.
In addition, when subject attrition is efficient, patterns might mirror the changing idea of the
example, maybe the elements of the marvels being scrutinized (Itzhak and Arthur,2007). Other
bewildering impacts might be made by verifiable occasions and changes in measures or recording
practices of factors over the long haul (Itzhak and Arthur,2007). For example, the dataset shows
that these data can reflect in one period.
Recommendations
One of the most efficient methods to solve this problem is to advocate the public to implement
epidemic prevention strategies such as increasing the publicity effort of epidemic items, discount
for epidemic items, providing online delivery, and focusing on employee infection. The first
strategy is that the retail market needs publicity to increase total sales during the epidemic. The
forecast model of facemask demonstrates that the number of facemasks sold is down from seasons
4 to 10. One reason is maybe citizens lack protection consciousness. Another reason is that people
intend to stay at home instead of buying goods outside to prevent infection.
Nevertheless, there is no good way in the long term, and the face mask can provide freedom to the
public to go out. Therefore, it is suggested that the retail market needs advertisements to remind
individuals about the importance of facemasks. The advertisements format can be including
pictures, videos, and messages.
Next, according to the regression between essential goods and confirmed, the sales of essential
goods would decrease due to increasing the number of confirmed cases. A reason might be that
people would not intend to go out frequently due to COVID-19 because most people choose to
purchase goods when they do not have enough stock. However, retail marketing should provide
the discount to customers at this period, for the public would intend to increase their store because
of discount.
Finally, the diagram above shows that the total sale decreases with the increase of confirmed cases.
Therefore, it is recommended that the retail market give more opportunities to transfer some of the
physical offline business patterns into online sales. The retail market can apply to the government
for delivery, which prevents businesses from having no or low income during the lockdown and
satisfies people’s demand for consumption. Developing a new clicks-and-mortar business model
that combines online e-commerce with physical retail outlets may solve the inconvenience of
residents not getting access to physical stores. One thing that needs to be noted is that businesses
need to take preventive measures to avoid a second epidemic outbreak because they must contact
many customers. Specifically, it is required to wear a facemask and show the health QR code
Assessment 4: Group Project Report (30%) ---- Group 1
14
before getting into the store. All the goods and productions need to be disinfected daily by the
staff.
Reflection
In the process of analysis, we took in the fundamental examination and factual apparatuses for
getting clients, markets, contenders, and ourselves. Apply accurately and information examination
abilities to genuine enormous informational indexes, zeroing in on applications instead of
techniques. We figured out how to depict, gather, and examine business information from various
fields like money, advertising, the executives, and hazard examination. Information experiences
and ideas would then be viably imparted to non-specialized crowds, such as data visualization. All
through the abilities above, it can help we are applying to genuine cases. Like this project, we are
utilizing various models to demonstrate our presumptions and give conclusions.
Appendix
Chart 1: Line Chart for Test and Confirmed
library(readxl)
MyData <- read_excel("C:/Users/Linda/Desktop/MyData.xlsx")
View(MyData)
library(ggplot2)
ggplot()+
geom_line(data=MyData,mapping=aes(x=Date,y=Test),color="blue")+
geom_line(data=MyData,mapping=aes(x=Date,y=Confirmed),color="red")
+
labs(title="Linechart for Test and Confirmed",y = "Daily number of
Test and confirmed cases")
Chart 2: Bubble Plot for the Sales of Essentialgoods
library(readxl)
myData <- read_excel("C:/Users/admin/Desktop/1190 Group Work/myData.x
lsx")
View(myData)
plot(myData$Confirmed~myData$Test)
symbols(myData$Confirmed~myData$Test,circles = myData$Essentialgoods,
inches = 0.4,bg="red",main = "A Bubble plot for sales of essential go
ods",xlab="Test", ylab="Confirmed" )
abline(lm(myData$Confirmed~myData$Test),col='black')
Table 1: Regression for Medicine and Confirmed
library(readxl)
myData <- read_excel("C:/Users/admin/Desktop/1190 Group Work/myData.x
Assessment 4: Group Project Report (30%) ---- Group 1
15
lsx")
View(myData)
Model=lm(Medicine~Confirmed,data=myData)
summary(Model)
##
## Call:
## lm(formula = Medicine ~ Confirmed, data = myData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -374.08 -51.80 1.25 57.14 118.77
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.313e+03 7.163e+00 183.235 < 2e-16 ***
## Confirmed 1.437e-03 3.808e-04 3.773 0.00018 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 74.08 on 516 degrees of freedom
## Multiple R-squared: 0.02685, Adjusted R-squared: 0.02496
## F-statistic: 14.23 on 1 and 516 DF, p-value: 0.0001801
summary(myData$Confirmed)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1033 10722 15574 16756 21731 39542
predict(Model,data.frame(Confirmed=15574))
## 1
## 1334.977
Table 2: Regression for Totalsales and Confirmed
library(readxl)
myData <- read_excel("C:/Users/jingy/OneDrive/words/myData.xlsx")
View(myData)
Model=lm(Totalsales~Confirmed,data=myData)
summary(Model)
##
## Call:
## lm(formula = Totalsales ~ Confirmed, data = myData)
##
## Residuals:
Assessment 4: Group Project Report (30%) ---- Group 1
16
## Min 1Q Median 3Q Max
## -64943 -42352 -29480 -11085 1918457
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 72229.4047 14613.3325 4.943 1.04e-06 ***
## Confirmed -1.6246 0.7769 -2.091 0.037 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 151100 on 516 degrees of freedom
## Multiple R-squared: 0.008403, Adjusted R-squared: 0.006482
## F-statistic: 4.373 on 1 and 516 DF, p-value: 0.037
summary(myData$Confirmed)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1033 10722 15574 16756 21731 39542
predict(Model,data.frame(Confirmed=15574))
## 1
## 46928.1
Table 3: Regression for Essentialgoods and Confirmed
ibrary(readxl)
myData <- read_excel("C:/Users/Orb/Desktop/myData.xlsx")
View(myData)
packages.required <- c("data.table", "ggplot2","magrittr","dplyr")
new.packages <- packages.required[!(packages.required %in% installed.
packages()[,"Package"])]
library(readxl)
myData <- read_excel("C:/Users/Orb/Desktop/myData.xlsx")
View(myData)
Model=lm(Essentialgoods~Confirmed,data=myData)
summary(Model)
##
## Call:
## lm(formula = Essentialgoods ~ Confirmed, data = myData)
##
## Residuals:
## Min 1Q Median 3Q Max
## -64471 -42048 -29401 -11151 1914684
##
Assessment 4: Group Project Report (30%) ---- Group 1
17
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 71483.7186 14561.5344 4.909 1.23e-06 ***
## Confirmed -1.6217 0.7741 -2.095 0.0367 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 150600 on 516 degrees of freedom
## Multiple R-squared: 0.008433, Adjusted R-squared: 0.006511
## F-statistic: 4.388 on 1 and 516 DF, p-value: 0.03667
summary(myData$Confirmed)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1033 10722 15574 16756 21731 39542
Table 4: Forecast for Facemasks
library(readxl)
MyData <- read_excel("C:/Users/Linda/Desktop/MyData.xlsx")
View(MyData)
library(forecast)
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
head(MyData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
##
dbl>
## 1 2020-01-30 00:00:00 1260 1033. 106 1989008. 198
4492.
## 2 2020-01-31 00:00:00 1386 1137. 81 749637. 74
6415
## 3 2020-02-01 00:00:00 1525. 1250. 60 1600316. 159
2408.
## 4 2020-02-02 00:00:00 1677. 1375. 101 171372. 16
6895.
## 5 2020-02-03 00:00:00 1845. 1513. 49 6079.
5700.
## 6 2020-02-04 00:00:00 2029. 1664. 51 1000644. 99
8389.
Assessment 4: Group Project Report (30%) ---- Group 1
18
## # ... with 6 more variables: Handsanitizer
## # Toilettissue
g
tail(MyData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
##
dbl>
## 1 2021-06-25 00:00:00 8833. 7420. 34 4012.
3518.
## 2 2021-06-26 00:00:00 8568. 7197. 36 25194. 2
4746.
## 3 2021-06-27 00:00:00 8311. 6981. 9 57637. 5
7031.
## 4 2021-06-28 00:00:00 8062. 6772. 31 11709. 1
0726.
## 5 2021-06-29 00:00:00 7820. 6569. 14 12821. 1
2459.
## 6 2021-06-30 00:00:00 7585. 6372. 21 58847. 5
8082.
## # ... with 6 more variables: Handsanitizer
## # Toilettissue
g
newData=ts(MyData$Facemask,start=c(2020,1,30),end=c(2021,6,30),frequ
ency=12)
TSReg=tslm(newData~trend+season)
summary(TSReg)
##
## Call:
## tslm(formula = newData ~ trend + season)
##
## Residuals:
## [1] -1.690e+02 -2.159e+02 -6.213e+02 1.131e+03 -4.210e+01 -8.263
e+01
## [7] -6.395e-14 -6.395e-14 -1.421e-14 -4.263e-14 -1.208e-13 6.395
e-14
## [13] 1.690e+02 2.159e+02 6.213e+02 -1.131e+03 4.210e+01 8.263
e+01
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
Assessment 4: Group Project Report (30%) ---- Group 1
19
## (Intercept) 2396.95 655.06 3.659 0.0146 *
## trend 128.08 40.24 3.183 0.0245 *
## season2 36.60 837.34 0.044 0.9668
## season3 437.56 840.23 0.521 0.6248
## season4 -1312.86 845.04 -1.554 0.1810
## season5 -131.52 851.72 -0.154 0.8833
## season6 -75.89 860.23 -0.088 0.9331
## season7 -136.25 1024.34 -0.133 0.8994
## season8 -106.47 1025.13 -0.104 0.9213
## season9 -68.79 1027.50 -0.067 0.9492
## season10 -22.83 1031.43 -0.022 0.9832
## season11 31.83 1036.91 0.031 0.9767
## season12 95.64 1043.91 0.092 0.9306
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 836.4 on 5 degrees of freedom
## Multiple R-squared: 0.7606, Adjusted R-squared: 0.186
## F-statistic: 1.324 on 12 and 5 DF, p-value: 0.4018
forecast(TSReg,h=4)
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jul 2021 4694.232 2808.674 6579.791 1410.111 7978.354
## Aug 2021 4852.095 2966.537 6737.654 1567.974 8136.217
## Sep 2021 5017.852 3132.293 6903.410 1733.730 8301.973
## Oct 2021 5191.896 3306.338 7077.455 1907.775 8476.018
plot(forecast(TSReg,h=4),xlab='Date',ylab='Facemask',main='Time Seri
es for Facemask')
Table 5: Forecast for Handsanitizer
library(readxl)
myData <- read_excel("C:/Users/Orb/Desktop/myData.xlsx")
View(myData)
head(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
##
dbl>
## 1 2020-01-30 00:00:00 1260 1033. 106 1989008. 198
4492.
## 2 2020-01-31 00:00:00 1386 1137. 81 749637. 74
Assessment 4: Group Project Report (30%) ---- Group 1
20
6415
## 3 2020-02-01 00:00:00 1525. 1250. 60 1600316. 159
2408.
## 4 2020-02-02 00:00:00 1677. 1375. 101 171372. 16
6895.
## 5 2020-02-03 00:00:00 1845. 1513. 49 6079.
5700.
## 6 2020-02-04 00:00:00 2029. 1664. 51 1000644. 99
8389.
## # ... with 6 more variables: Handsanitizer
## # Toilettissue
g
tail(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
##
dbl>
## 1 2021-06-25 00:00:00 8833. 7420. 34 4012.
3518.
## 2 2021-06-26 00:00:00 8568. 7197. 36 25194. 2
4746.
## 3 2021-06-27 00:00:00 8311. 6981. 9 57637. 5
7031.
## 4 2021-06-28 00:00:00 8062. 6772. 31 11709. 1
0726.
## 5 2021-06-29 00:00:00 7820. 6569. 14 12821. 1
2459.
## 6 2021-06-30 00:00:00 7585. 6372. 21 58847. 5
8082.
## # ... with 6 more variables: Handsanitizer
## # Toilettissue
g
library(forecast)
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
newData=ts(myData$Handsanitizer,start=c(2020,01,30),end=c(2021,06,3
0),frequency=12)
TSReg=tslm(newData~trend+season)
summary(TSReg)
Assessment 4: Group Project Report (30%) ---- Group 1
21
##
## Call:
## tslm(formula = newData ~ trend + season)
##
## Residuals:
## [1] 1.543e+02 3.563e+01 8.776e-01 -3.256e+01 -6.431e+01 -9.395
e+01
## [7] -1.066e-14 -1.066e-14 1.066e-14 -2.132e-14 1.421e-14 0.000
e+00
## [13] -1.543e+02 -3.563e+01 -8.776e-01 3.256e+01 6.431e+01 9.395
e+01
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 732.166 97.954 7.475 0.000677 ***
## trend 223.523 6.017 37.147 2.66e-07 ***
## season2 6.160 125.211 0.049 0.962669
## season3 -60.513 125.644 -0.482 0.650416
## season4 -116.290 126.363 -0.920 0.399646
## season5 -160.323 127.362 -1.259 0.263676
## season6 -191.687 128.635 -1.490 0.196367
## season7 -330.395 153.175 -2.157 0.083505 .
## season8 -357.275 153.293 -2.331 0.067156 .
## season9 -364.490 153.647 -2.372 0.063774 .
## season10 -350.075 154.235 -2.270 0.072461 .
## season11 -311.866 155.054 -2.011 0.100475
## season12 -247.484 156.102 -1.585 0.173732
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 125.1 on 5 degrees of freedom
## Multiple R-squared: 0.9967, Adjusted R-squared: 0.9889
## F-statistic: 127.7 on 12 and 5 DF, p-value: 2.099e-05
forecast(TSReg,h=4)
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jul 2021 4648.710 4366.753 4930.667 4157.619 5139.801
## Aug 2021 4845.353 4563.396 5127.310 4354.262 5336.444
## Sep 2021 5061.661 4779.704 5343.618 4570.570 5552.752
## Oct 2021 5299.599 5017.642 5581.556 4808.508 5790.690
plot(forecast(TSReg,h=4),xlab='Date',ylab='Handsanitizer',main='Time
Series for Handsanitizer')
Assessment 4: Group Project Report (30%) ---- Group 1
22
Table 6: Forecast for Totalsales
library(readxl)
myData <- read_excel("1190 Group Work/myData.xlsx")
View(myData)
head(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
##
dbl>
## 1 2020-01-30 00:00:00 1260 1033. 106 1989008. 198
4492.
## 2 2020-01-31 00:00:00 1386 1137. 81 749637. 74
6415
## 3 2020-02-01 00:00:00 1525. 1250. 60 1600316. 159
2408.
## 4 2020-02-02 00:00:00 1677. 1375. 101 171372. 16
6895.
## 5 2020-02-03 00:00:00 1845. 1513. 49 6079.
5700.
## 6 2020-02-04 00:00:00 2029. 1664. 51 1000644. 99
8389.
## # ... with 6 more variables: Handsanitizer
## # Toilettissue
g
tail(myData)
## # A tibble: 6 x 12
## Date Test Confirmed Recovered Totalsales Essentia
lgoods
##
dbl>
## 1 2021-06-25 00:00:00 8833. 7420. 34 4012.
3518.
## 2 2021-06-26 00:00:00 8568. 7197. 36 25194. 2
4746.
## 3 2021-06-27 00:00:00 8311. 6981. 9 57637. 5
7031.
## 4 2021-06-28 00:00:00 8062. 6772. 31 11709. 1
0726.
## 5 2021-06-29 00:00:00 7820. 6569. 14 12821. 1
2459.
## 6 2021-06-30 00:00:00 7585. 6372. 21 58847. 5
Assessment 4: Group Project Report (30%) ---- Group 1
23
8082.
## # ... with 6 more variables: Handsanitizer
## # Toilettissue
g
library(forecast)
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoo
newData=ts(myData$Totalsales,start=c(2020,01,30),end=c(2021,06,30),f
requency=12)
TSReg=tslm(newData~trend+season)
summary(TSReg)
##
## Call:
## tslm(formula = newData ~ trend + season)
##
## Residuals:
## [1] 5.844e+05 -3.881e+04 3.024e+05 -4.707e+05 -4.043e+05 2.704
e+04
## [7] 4.366e-11 1.455e-11 2.910e-11 1.455e-11 7.276e-11 5.821
e-11
## [13] -5.844e+05 3.881e+04 -3.024e+05 4.707e+05 4.043e+05 -2.704
e+04
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1472315 448610 3.282 0.0219 *
## trend -67702 27558 -2.457 0.0575 .
## season2 -548464 573442 -0.956 0.3828
## season3 28739 575425 0.050 0.9621
## season4 -559460 578715 -0.967 0.3781
## season5 -723405 583290 -1.240 0.2699
## season6 -92496 589120 -0.157 0.8814
## season7 -992727 701509 -1.415 0.2162
## season8 -877908 702050 -1.250 0.2665
## season9 -838231 703670 -1.191 0.2870
## season10 -582707 706363 -0.825 0.4470
## season11 -724656 710116 -1.020 0.3543
## season12 -533442 714913 -0.746 0.4891
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
Assessment 4: Group Project Report (30%) ---- Group 1
24
## Residual standard error: 572800 on 5 degrees of freedom
## Multiple R-squared: 0.7223, Adjusted R-squared: 0.05591
## F-statistic: 1.084 on 12 and 5 DF, p-value: 0.5007
forecast(TSReg,h=4)
## Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
## Jul 2021 -806755.2 -2098058 484547.3 -3055847 1442337
## Aug 2021 -759639.3 -2050942 531663.2 -3008731 1489452
## Sep 2021 -787664.6 -2078967 503637.9 -3036756 1461427
## Oct 2021 -599842.7 -1891145 691459.8 -2848935 1649249
plot(forecast(TSReg,h=4),xlab='Date',ylab='Totalsales',main='Time Se
ries for Totalsales')
Reference
[1] Eric, T, B., Manish,G., Praveen,K., Sudhir, Voleti. (March 2017). The Role of Big Data and
Predictive Analytics in Retailing. Volume 93, Issue 1. Available from
[2] Dekimpe, Marnik G. and Dominique M. Hanssens (2000), “Time-series Models in Marketing:
Past, Present and Future,” International Journal of Research in Marketing, 17 (2), 183–93.
[3] Ailawadi, Kusum L., Donald R. Lehmann and Scott A. Neslin (2003), “Revenue Premium as
an Outcome Measure of Brand Equity,” Journal of Marketing, 67 (4), 1–17.
[4] Larson, Jeffrey S., Eric T. Bradlow and Peter S. Fader (2005), “An Exploratory Look at
Supermarket Shopping Paths,” International Journal of Research in Marketing, 22 (4), 395–414.
[5] Ritchie, H., Ortiz-Ospina, E., Beltekian, D., Mathieu, E., Hasell, J., Macdonald, B., Giattino,
C., Appel, C., Rodés-Guirao, L., & Roser, M. (March 5, 2020). Coronavirus (covid-19)
VACCINATIONS - statistics and research. Our World in Data. Available from
[6] Evans, A., Santos, E., Ford, M. (May 1, 2020). Impact of COVID-19 on Vitamins, Minerals
and Supplements. Volume XXII, issue 32. Available from
