UNIVERSITY OF ECONOMICS HO CHI MINH CITY

College of Business
School of International Business and Marketing

CASE STUDY REPORT

on
APPLIED STATISTICS

Case Study: ANALYZING CUSTOMER SATISFACTION

BASED ON KNOWLEDGE FROM CHAPTERS 2-3

Instructor: Dr. Chu Nguyen Mong Ngoc

Class: DH51MRF001
Group members:
Tran Linh Thuc Anh 31251021709
Quan Hai Khang Duy 31251020047
Vo Hoang Nam 31251022217
Tong Phuong Nhi 31251020980
Hoang Kim Tu 31251021382
 1

TEAM MEMBER EVALUATION FORM

Full name Student ID Contribution Evaluation (%)

Trần Linh Thục Anh 31251021709 100%

Quan Hải Khang Duy 31251020047 100%

Võ Hoàng Nam 31251022217 100%

Tống Phương Nhi 31251020980 100%

Hoàng Kim Tú 31251021382 100%

Ho Chi Minh, November 25th 2025

 2

case
1. INTRODUCTION...........................................................................................................................................
1.1. Case study introduction............................................................................................................................
1.2. Research objective...................................................................................................................................
2. PRELIMINARY ANALYSIS........................................................................................................................
2.1. Reason for preliminary analysis...............................................................................................................
2.2. Scale identification of data.......................................................................................................................
2.3. Data quality assessment...........................................................................................................................
2.3.1. Completeness check........................................................................................................................
2.3.2. Validity and range check.................................................................................................................
2.3.3. Logical consistency check..............................................................................................................
3. DESCRIPTIVE STATISTICAL ANALYSIS OF CUSTOMER DEMOGRAPHICS.............................
3.1. Age...........................................................................................................................................................
3.2. Gender......................................................................................................................................................
To visualize the gender data, we use a bar chart which shows the respondent’ gender:................................
3.3. Educational Level....................................................................................................................................
3.4. Monthly income.......................................................................................................................................
3.5. Conclusion...............................................................................................................................................
4. SATISFACTION ANALYSIS.......................................................................................................................
4.1. Satisfaction of each product.....................................................................................................................
4.1.1. Descriptive statistics for satisfaction in general..............................................................................
4.1.2. Comparison across three products..................................................................................................
The comparative analysis of satisfaction scores shows a difference in the three
product lines:.............................................................................................................................................
4.1.3. Conclusion......................................................................................................................................
4.2. Analyzing factors affecting satisfaction...................................................................................................
4.2.1. Method............................................................................................................................................
[Link]. Analysis of Categorical Variables..........................................................................................
[Link]. Analysis of Quantitative Variables........................................................................................
[Link]. Supporting Descriptive Procedures........................................................................................
4.2.2. Relationship between Satisfaction and Demographic factors of ultrasonic
insect-repelling vacuum robots (Product 2)..............................................................................................
[Link]. Usage frequency.....................................................................................................................
[Link] Gender.....................................................................................................................................
[Link]. Age.........................................................................................................................................
[Link]. Education level.......................................................................................................................
[Link]. Income....................................................................................................................................
[Link]. Conclusion.............................................................................................................................
4.2.3. Relationship between Satisfaction and Demographic factors of Smart air
purifiers (Product 1)..................................................................................................................................
 3
[Link]. Usage frequency.....................................................................................................................
[Link] Gender.....................................................................................................................................
[Link]. Age.........................................................................................................................................
[Link]. Education Level.....................................................................................................................
[Link]. Income....................................................................................................................................
[Link]. Conclusion.............................................................................................................................
4.2.4. Relationship between Satisfaction and Demographic factors of fan integrating
room deodorizing function (Product 3).....................................................................................................
[Link]. Usage frequency...........................................................................................................................
[Link]. Gender....................................................................................................................................
[Link]. Age.........................................................................................................................................
[Link]. Education level.......................................................................................................................
[Link]. Income....................................................................................................................................
[Link]. Conclusion.............................................................................................................................
5. CONCLUSION: DISCUSSION AND RECOMMENDATION.................................................................
7. ACKNOWLEDGEMENT..............................................................................................................................
 4

1. INTRODUCTION

1.1. Case study introduction

Case study: Analyzing customer satisfaction based on knowledge from chapter 2, 3 of the
book “Statistics in Business and Economics”.

Context: Te Jia Company specializes in selling three product lines in a highly competitive
market, where many competitors are constantly improving quality and prices to attract
customers. The market today is witnessing rapid changes in consumer trends, with customers
increasingly prioritizing high-quality products, reasonable prices, and good after-sales
service. Additionally, customer preferences are becoming more diverse, especially the
demand for environmentally friendly products and those integrated with modern technology.
Some of the standout products in Te Jia’s market include smart air purifiers (1), ultrasonic
insect-repelling vacuum robots (2), and fan integrating room deodorizing function (3).
To maintain its market position, Te Jia has conducted a customer survey to evaluate customer
satisfaction and product usage behavior, in order to provide recommendations for improving
product quality and marketing strategies. The collected data includes: age, gender, education
level, monthly income, satisfaction ratings for products, and usage frequency for each
product.

1.2. Research objective

- Describe customer profiles based on age, gender, education level, and income.
- Assess and compare satisfaction levels for each product line.
- Analyze the relationship between satisfaction levels and usage frequency.
- Analyze the relationship between satisfaction levels and demographic variables.
- Draw conclusions and recommendations.

2. PRELIMINARY ANALYSIS

2.1. Reason for preliminary analysis

Preliminary analysis is a crucial step before proceeding any in-depth statistical analysis, of
which purposes are to ensure data validity while facilitating the selection of appropriate
analytical methods. Specifically, in the subsequent parts, our team aims to achieve these key
objectives:

- Scale identification: By identifying the correct measurement scale for each variable
(Nominal, Ordinal, Interval, or Ratio), we would be able to select suitable descriptive
statistical tools.
 5
- Data quality assessment: This step is to detect possible issues such as missing
values, data entry errors, or outliers in numeral variables like Age, Monthly Income,
and Usage Frequency, as well as logic errors in categorical ones like Gender or
Education Level.

2.2. Scale identification of data

Identifying the measurement scales of all variables helps ensure that each type of data is
analyzed with suitable statistical methods. This part clarifies what operations are appropriate
for each variable before moving into further analysis.

Variable Description Data type Measurement scale

Age Customer age (19-64) Quantitative Ratio

Gender Male/Female Categorical Nominal

Likert scale (1 = very Categorical (treated

Satisfaction Score
dissatisfied → 5 = very as quantitative data Ordinal
(Product 1-3)
satisfied) for analysis)

Times per week

Usage Frequency
customers use each Quantitative Ratio
(Product 1-3)
product

Primary →
Education Level Categorical Ordinal
Postgraduate

Monthly salary
Monthly Income
reported by Quantitative Ratio
(VND)
respondents
Table 2.2. Scale identification of data

Note: Satisfaction Scores (Product 1-3) are ordinal categorical data, but they are treated as
quantitative in this report. This approach is a common practice in applied survey analysis,
where Likert-scale responses are handled as numerical values to help summarize trends more
clearly, without assuming that each step on the scale is exactly the same distance.

2.3. Data quality assessment

This part of assessment is to ensure validity, accuracy, and reliability of the surveyed data. It
focused on three primary criteria: completeness, validity against established ranges, and
logical consistency.

2.3.1. Completeness check

 6
For completeness check, our team examined all the data field for missing values, specifically
cells that are either “#N/A” or blank. This observation revealed that the collected data is fully
complete, with 0% of missing values detected across all the variables, thus showing sufficient
observations are available for analysis and eliminating the requirement of removal or
advanced tools like imputation.

2.3.2. Validity and range check

Data validity was confirmed by checking that all recorded values adhered to the pre-defined
theoretical ranges established. The analysis confirmed that no out-of-range values were
detected across the dataset. Specifically:

- Categorical variables: All entries for Gender and Education Level were correctly
mapped to the defined categories.
- Continuous variables: Values for Age were confirmed to be between 19 and 64,
Monthly Income was confirmed to be between 4 million VND and 18 million VND,
Usage Frequency across three products was confirmed to be between 0 and 14 times
per week, and all Satisfaction Scores were strictly within the 1-to-5 Likert scale.

These findings confirm the high accuracy of the data recording process, establishing a
reliable foundation for all numerical and statistical analyses to follow.

2.3.3. Logical consistency check

This analysis verified the crucial logical rule. For this study, it is assumed that customers
should only provide satisfaction ratings for the product they have actually used (i.e., Usage
Frequency = 0 should logically preclude a Satisfaction Score). This assumption reflects the
analytical design of the survey, where Satisfaction Score represents real usage experience
rather than expectations.

The check identified 17 cases of inconsistency, in which respondents rated products they
reported not using. Notably, one respondent (Customer 25) did so for Product 1 and 2,
resulting in 17 inconsistencies across 16 respondents out of the 50 surveyed customers (32%).

The cases are as follows:

- Product 1: Customers 2, 7, 14, 20, 25, 27, 34, 41, 46

- Product 2: Customers 4, 24, 25, 39
- Product 3: Customers 10, 29, 30, 38

These findings indicate a significant logical inconsistency within the dataset. As these
satisfaction scores are unreliable and invalid, their existence must be carefully acknowledged
as they introduce potential bias into the assessment and comparison of customer satisfaction
levels.
 7
3. DESCRIPTIVE STATISTICAL ANALYSIS OF CUSTOMER
DEMOGRAPHICS

3.1. Age

To visualize the age data, we input a boxplot below that illustrates the distribution of
respondents’ ages, allowing us to observe the central tendency, variability, and potential
outliers within the dataset.

Chart 3.1. Box Plot of customers’ age

The data shows a wide age distribution ranging from 19 to 64 years old, with no statistical
outliers. The group centers around a median age of 40, with the majority of respondents (the
middle 50%) falling between 31 and 50 years old.

3.2. Gender

To visualize the gender data, we use a bar chart which shows the respondent’ gender:

Chart 3.1. Bar Chart of customers’ gender

 8

The data reveals a significant difference between two genders. The number of female
respondents is as twice as the number of male respondents (34 female respondents vs.
16 male respondents).

3.3. Educational Level

To visualize the distribution of the consumers' educational backgrounds, we will present a bar
chart for the Education Level variables.

Figure 3.1. Bar Chart of customers’ Education Levels

The bar chart highlights that the customer base is mainly composed of individuals with
intermediate education, where High School (nhs = 16) and College (nc = 14) graduates
collectively account for 60% of the total sample. While there is a notable segment of higher
education, with University and Postgraduate degrees making up 36% of respondents, the
Primary School group is negligible at only 4%.

3.4. Monthly income

To analyze the variability and distribution shape of the respondents' earnings, we employ a
histogram.
A histogram is selected as the primary visualization tool because Monthly Income is a
quantitative variable measured on a ratio scale.
 9
Figure 3.4. Histogram for consumers’ monthly income

The histogram illustrates the distribution of customer monthly income, which begins in the 4–
6 million VND range and rises steadily. The data is heavily concentrated in the middle, as the
10–12 million VND and 13–15 million VND intervals each contain 14 customers,
collectively accounting for 56% of the total sample. In contrast, the highest income bracket of
16–18 million VND represents a small minority with only 3 customers, confirming that the
majority of the respondents earn between 10 and 15 million VND.

3.5. Conclusion
In conclusion, the data reveals that the target audience is mainly female, middle-aged, and
firmly situated within the middle-income class.

The demographic includes:

• Gender & Age: The customer base is significantly skewed toward females, who
outnumber male respondents by a ratio of approximately 2:1. The age distribution
reveals a mature audience, with a median age of 40 years and the majority (50%)
falling between the ages of 31 and 50. This suggests a consumer group that is likely
established in their careers and family lives.
• Education Level: The survey shows a moderate-to-high level of education, with 60% of
respondents holding High School or College degrees, and a further 36% possessing
University or Postgraduate qualifications.
• Monthly Income: The majority of respondents fall into the 8 million VND – 14 million
VND, with the most significant earning between 10 million and 12 million VND.

4. SATISFACTION ANALYSIS

4.1. Satisfaction of each product

4.1.1. Descriptive statistics for satisfaction in general

This section presents the descriptive statistics for customer satisfaction collected from the
survey. To provide an initial overview of the data, we summarized the number of respondents
who selected each satisfaction level for the three products. Table [Link]. reports the
frequency distribution of satisfaction ratings, illustrating how customer responses are
distributed across the five-point Likert scale for each product.

Satisfaction Score Product 1 Product 2 Product 3

1 11 9 8
2 4 17 11
3 11 7 5
 10
4 12 10 11
5 12 7 15
Total 50 50 50
Table [Link]. Frequency distribution of each satisfaction level across three products

As is shown by Table [Link]., Product 1 has a relatively even spread of responses across all
satisfaction levels, except for level 2. Meanwhile, Product 2 receives more ratings at level 2
and 4. Product 3 appears to perform better overall, with a concentration of higher ratings,
especially at level 5.

However, to comprehensively assess the customer satisfaction across three products, it is

necessary to take other indicators such as mean, mode and median into consideration. Table
[Link]. below summarizes those three indicators calculated for each product’s statistics,
assuming that a higher mode or a higher mean indicate higher satisfaction, while the standard
deviation indicates the spread or variability of the responses.

Indicator Product 1 Product 2 Product 3

Mean 3,2 2,78 3,28
Mode 4 2 5
Standard deviation 1,470804306 1,344527836 1,498843091
Table [Link]. Mean, Mode, and Standard deviation of three products’ Satisfaction Score

4.1.2. Comparison across three products

The comparative analysis of satisfaction scores shows a difference in the three product lines:

1. Product 1 demonstrates a balanced performance, with a mean score of 3.20 and a

relatively even distribution across the scale. While customer satisfaction with Product
1 is lower compared to Product 3, it remains superior to Product 2.
2. In contrast, Product 2 exhibits a significantly lower average score (Mean = 2.78) than
the other two products. Notably, a substantial portion of users (34%) rated it at level
2, indicating potential issues with product quality or market alignment that necessitate
immediate attention.
3. Product 3 stands out as the product with the highest average satisfaction score (Mean
= 3.28) and the highest median score (Median = 4.0). Approximately 30% of
respondents provide it with a perfect rating, which is notably higher than the ratings
for Product 1 and Product 2.

4.1.3. Conclusion

In conclusion, the analysis reveals a clear order in customer satisfaction levels among the
three products. Product 3 stands out as the market leader, boasting the highest mean
satisfaction score of 3.28 and the most positive customer sentiment, with a mode of 5.
 11
Product 1 maintains a stable but somewhat polarized position, serving as a mid-tier offering
with average performance. On the other hand, Product 2 is clearly the underperformer, with
the lowest mean score of 2.78 and a mode of only 2. This indicates significant challenges in
terms of customer acceptance. While Product 3 should be considered a benchmark for
success, Product 2 requires immediate strategic intervention to address its quality or market
fit issues.
Having established the demographic profile of the respondents: gender, age, education, and
monthly income, the next step is to confirm whether these characteristics have an influence
on the perception of the product. Therefore, we shall proceed to Section 4.2, analyzing the
correlation between these demographic factors and customer satisfaction. Utilizing statistical
correlation and hypothesis testing, we will point out the existence of any significant
difference among user groups.

4.2. Analyzing factors affecting satisfaction

4.2.1. Method

Before presenting the analytical results for all three products, it is necessary to outline the
statistical methods used throughout the satisfaction analysis. This section describes the
techniques applied to examine how demographic characteristics and usage patterns relate to
customer satisfaction across the dataset. Explaining these methods and the rationale for using
them helps our team to ensure that the results presented later are clear and easier to interpret.

[Link]. Analysis of Categorical Variables

For demographic factors that are reported as categorical data, such as Gender and Education
Level, two tabular and visual techniques are being used:
● Crosstabulation: summarises the distribution of Satisfaction Score within each
demographic group.
● Percentage distribution: converts raw frequencies into comparable proportions,
allowing differences between groups to be interpreted fairly even when group sizes
are not equal.
● Side-by-side bar charts: used to visualise how satisfaction ratings vary between
categories, providing a clear comparison of response patterns.

[Link]. Analysis of Quantitative Variables

For quantitative variables including Age, Monthly Income and Usage Frequency, the analysis
applied methods suited to identifying numerical and relational patterns:
● Box plot: illustrates the distribution, variability and potential outliers of continuous
variables across satisfaction groups, supporting comparisons that are less sensitive to
skewness.
● Correlation coefficient: quantifies the strength and direction of these relationships,
allowing numerical confirmation of any patterns observed in the scatter plots.
● Histogram: illustrates how continuous variables are distributed, helping identify
skewness, concentration patterns, and unusual gaps that support later interpretation.
 12
● Scatter diagram: provides a graphical view of how satisfaction varies with each
continuous predictor and helps assess visible linear trends or irregularities.
[Link]. Supporting Descriptive Procedures
Across both categorical and continuous variables, descriptive summaries, such as comparing
means, frequency counts and distributional patterns are being used to provide context and to
support the interpretation of the analytical results.
These statistical techniques establish a consistent and structured framework for analysing
how demographic characteristics and usage behaviour relate to customer satisfaction across
all three products.

4.2.2. Relationship between Satisfaction and Demographic factors of

ultrasonic insect-repelling vacuum robots (Product 2)

[Link]. Usage frequency

To examine the relationship between the satisfaction of users with the Usage Frequency, we
use an Excel function to compute the correlation coefficient and a scatter diagram to have a
visualized view of the trend. Additionally, to gain deeper insights into behavioral differences,
we calculated the mean Usage Frequency for each specific Likert satisfaction rating.
a) Correlation coefficient (Excel function)
The correlation coefficient turns out to be rs= 0.3263303
b) Scatter diagram
Here is the scatter diagram with the vertical axis appearing as the levels of satisfaction and
the horizontal axis representing the frequency of Product 2 being used.

Figure [Link]. Scatter diagram of Satisfaction Score and Usage Frequency of Product 2

c) Mean of Usage Frequency in each Likert-scale’s satisfaction rating

Likert-scale’s Mean of Usage

satisfaction rating Frequency
 13

1 4

2 5.59

3 7.43

4 10

5 6.14

Table [Link]. Means of Usage Frequency in each Likert’s rating

The analysis indicates a moderately positive correlation between Usage Frequency and the
satisfaction level for Product 2, with the correlation coefficient being rs ≈ 0.33. As shown
in the scatter diagram, there is a visible upward trend: customers who use the robot vacuum
more often tend to give higher ratings. However, Usage Frequency notably drops for the
highest rating (5). This fluctuation does not change the general upward relationship but shows
that the heaviest usage does not necessarily correspond to the highest satisfaction. Overall,
Usage Frequency plays a moderate role in affecting customers’ satisfaction of Product
2.

[Link] Gender
We determine the effect of genders on the satisfaction level by arranging both data into a
crosstabulation from the Pivot Table from Excel. And then based on the arranged data, we
generate a percentage comparison table between these two factors in the table below.

Gender Mean of Satisfaction Score

Male 2.563

Female 2.882
Table [Link]. Mean of Satisfaction Score of two genders

On average, female users report slightly higher satisfaction than male users, with mean
satisfaction scores of 2.882 and 2.563, respectively. Although the difference is not large, just
about 0.319, it suggests that female customers tend to evaluate the product more positively
than males.
However, this only reflects the difference in their proportions in rating. So we use a side-by-
side bar chart to visualize the relationship of Genders and Satisfaction Score in order to get a
clearer view on how customers of two genders participate in giving product assessment.
 14

Figure [Link]. Side-by-side bar chart for Gender and Satisfaction Score

Overall, both the mean comparison and the side-by-side bar chart show that Gender has only
a limited influence on satisfaction. While the distributions visualized in the chart show
some small variations between male and female respondents, these differences are not
substantial, and the average scores remain close, 2.563 for males and 2.882 for females. This
narrow gap suggests that gender does not meaningfully affect satisfaction, as both groups
report broadly similar evaluation levels.

[Link]. Age
To visualize the relationship between Age of users and Satisfaction Score, we use box plots
to determine the correlation of these two factors.
 15
Figure [Link]. Box-plot chart of Age and Satisfaction

The box-plot chart shows that age patterns vary across satisfaction levels. Lower satisfaction
levels such as 1 or 2 are associated with a wide age spread, including both younger and older
users, while higher satisfaction scores tend to come from older customers who are mostly in
their thirties or above, with score 5 having the highest proportion of medians.
As a result, this diagram suggests a mildly positive trend where older users appear more
likely to rate Product 2 with neutral or higher satisfaction scores, while younger users tend to
give lower evaluation. In short, older users are seen to be more satisfied with Product 2.
However, the variation within each group remains considerable, indicating that age
influences satisfaction only to a limited extent rather than being a strong determinant.

[Link]. Education level

As for a categorical variable like Education Level, we begin with arranging the data of both
Education Level and Satisfaction Score into a crosstabulation. With that data set, we continue
to compute the percentage of the arranged data of both factors. This results in the table below.

Grand Total
Satisfaction Score Product 2
Education Level 1 2 3 4 5
Primary school 0% 50% 0% 0% 50% 100%
High School 18.75% 31.25% 12.5% 37.5% 0% 100%
College 7.1429% 42.8571% 21.4286% 14.2857% 14.2857% 100%
University 20% 30% 20% 20% 10% 100%
Postgraduate 37.5% 25% 0% 0% 37.5% 100%
Table [Link]. Percentage demonstration of Education Level and Satisfaction Score of
Product 2
Table [Link]. illustrates how satisfaction scores for Product 2 differ across education levels.
College and University respondents tend to give moderate ratings, with approximately 43%
and 30% respectively rated at level 2. In contrast, both Primary school and Postgraduate
groups show highly polarized responses, concentrating at both ends of the scale. High School
respondents also display substantial variation. Specifically, 31.25% rated the product at level
2 and 37.5% rated at level 4. Overall, this suggests that the relationship of Education
Level and Satisfaction Score does not have a clear pattern, but higher education groups
appear more divided, while mid-education groups show more moderate satisfaction.

[Link]. Income
To evaluate how Monthly Income relates to their satisfaction level with Product 2, we
utilized an Excel function to compute the correlation coefficient and created a scatter diagram
to visually observe the distribution of income across satisfaction levels.
a) Correlation coefficient (Excel function)
The correlation coefficient turns out to be rs= -0.168952
b) Scatter diagram
 16
The scatter diagram below presents Monthly Income on the vertical axis and the Likert
Satisfaction Levels for Product 2 on the horizontal axis.

Figure [Link]. Scatter diagram showing graphical relationship between Monthly Income
and Satisfaction Score of Product 2.

The observed relationship between Monthly Income and the Satisfaction Score for Product 2
is negligible, as reflected by the weak negative correlation (r ≈ –0.17). It is also visualized in
the scatter diagram, indicating that income is not a decisive predictor of customer
satisfaction. Overall, the slight negative coefficient suggests a subtle tendency where higher-
income customers might be slightly more critical than lower-income ones. However,
satisfaction levels remain relatively independent of financial status.

[Link]. Conclusion
Based on the previous analysis, Usage Frequency and demographic characteristics have
varying degrees of association with customer satisfaction toward Product 2:
In the former, Usage Frequency demonstrates the clearest pattern, with a moderate positive
correlation indicating that the more frequent the users are, the higher satisfaction they report.
In the latter, Gender and Education Level show variations in distribution, but these mainly
imply response proportion rather than strong explanatory pattern. Age presents a mild upward
tendency, where older users appear more likely to provide neutral or higher ratings.
Meanwhile, users’ income shows almost no meaningful relationship with satisfaction,
suggesting that financial status is not a determinant of customers’ evaluation of the product.
Therefore, demographic factors only play a minor role in shaping customers’ evaluations of
Product 2.

4.2.3. Relationship between Satisfaction and Demographic factors of Smart

air purifiers (Product 1)
 17
[Link]. Usage frequency

We employed two methods to investigate the association between user satisfaction and usage
frequency of the smart air purifiers: calculating the correlation coefficient using the Excel
function =CORREL and generating a scatter diagram for graphical trend analysis.

Using the Microsoft Excel function, the correlation coefficient between Usage Frequency and
Satisfaction Score turns out to be rs= -0.077898, which is near 0, suggesting a no linear
relationship.

To visually confirm this negligibly weak relation and to explore possible influence of any
outliers, a scatter diagram was generated.

Figure [Link]. Scatter plot for Product 1’s usage frequency

The visualization of the data points confirms that there is no discernible pattern linking the
two variables, solidifying the conclusion that Usage Frequency is not a primary driver of
customer satisfaction.

[Link] Gender
The relationship between Gender and Customer Satisfaction was investigated using two
distinct methods to ensure a comprehensive assessment: a quantitative comparison of mean
satisfaction scores between two genders and a visual analysis of the full distribution profile.

Gender Mean Satisfaction Score

Male 3.125

Female 3.235
Table [Link]. Mean satisfaction score of two genders

As is shown by Table [Link]., both means still suggest an overall positive perception (above
the neutral score of 3). However, a difference of 0.110 points exists, indicating that females
are, on average, moderately more satisfied with Product 1 than males.
 18
According to Figure [Link]., there are clear differences in satisfaction levels between male
and female respondents. Female participants consistently report higher frequencies at scores 3
and 4, suggesting a stronger tendency toward moderate to high satisfaction. In contrast, male
responses appear more dispersed, with relatively high frequencies at both ends of the scale.
This pattern implies that female respondents generally show more stable and positive
satisfaction levels, whereas male respondents express a more varied range of satisfaction.

Figure [Link]. Side-by-side bar chart for Gender and Satisfaction score

Based on both the distribution chart and the mean satisfaction scores, Gender does not
appear to be a strong driver of satisfaction levels. While the bar chart shows that female
respondents tend to have higher frequencies at moderate-to-high satisfaction scores, the
difference in average satisfaction between genders is very small (0.11 points on a 5-point
scale). This finding suggests that although there are some visible variations in distribution,
the overall satisfaction levels between males and females are largely similar.

[Link]. Age

In this section, we use a box plot to visualize the relationship between consumers’ Ages and
their Satisfaction Score for smart air purifiers products.
 19

Figure [Link]. Box plot for relationship between Age and Satisfaction Score for Product 1

The box plot reveals distinct age patterns where lower satisfaction scores are concentrated
among older users, whereas higher scores (4 and 5) extend to include the youngest
demographic. This suggests that younger consumers are generally more satisfied with
Product 1 compared to middle-aged customers, who dominate the lowest rating. However,
the considerable variation across groups implies that age influences satisfaction only to a
limited extent rather than being a strong determinant.

[Link]. Education Level

To determine the effect of gender on satisfaction levels, we arranged the data into a
crosstabulation using an Excel Pivot Table. Based on this arranged data, we then generated a
percentage comparison table between these two factors.

Satisfaction Score Product 1

Education Level 1 2 3 4 5
Primary school 0% 0% 0% 8,33% 8,33%
High School 36,36% 50% 18,18% 41,67% 25%
College 27,27% 50% 36,36% 16,67% 25%
University 27,27% 0% 18,18% 8,33% 33,33%
Postgraduate 9,09% 0% 27,27% 25% 8,33%
Total 100% 100% 100% 100% 100%
 20
Table [Link]. Percentage Crosstabulation for the relationship between Satisfaction Score
and Education Level of Product 1

The crosstabulation highlights a polarized response from University graduates, who

constitute the largest portion of the most satisfied customers (Score 5) while also significantly
contributing to the most dissatisfied group. Similarly, High School graduates show highly
mixed experiences, dominating the 'Satisfied' tier (Score 4) but also leading in the lowest
ratings. To conclude, there appears to be no significant correlation between consumers'
education levels and their satisfaction scores for the smart air purifier, as the ratings are
widely distributed across all groups without a clear linear trend.

[Link]. Income
This analysis utilized two complementary methods: calculating the correlation coefficient to
quantify the strength and direction of the linear association between the two variables -
monthly income and satisfaction score, and generating a scatter diagram to visually assess
the data's distribution.
Using the Microsoft Excel function =CORREL, the correlation coefficient between income
and satisfaction score turns out to be rs= 0.03710434, which is near 0, suggesting a no linear
relationship.
In addition, a scatter plot was used to provide a visual assessment of the trend, allowing for
confirmation of the weak linear association.

Figure [Link]. Scatter plot for Monthly income and Product 1’s Satisfaction

This dispersed spread of data confirms the weak relationship between the two variables.
Therefore, Product 1’s customer satisfaction is unlikely to be influenced by Monthly
Income.

[Link]. Conclusion
 21
Based on the comprehensive analysis of the relationship between user satisfaction and
demographic factors for Product 1 (Smart Air Purifiers), we can conclude:

First, Usage Frequency and Monthly Income do not exhibit a significant link to customer
satisfaction. The scatter diagrams for both variables display nearly horizontal trend lines with
widely dispersed data points, indicating that neither a customer's financial status nor their
intensity of product usage acts as a strong determinant of their happiness. High and low
ratings are distributed evenly across all income brackets and usage levels.

Second, demographic characteristics including age, gender, and education reveal distinct
behavioral patterns. Regarding age, younger users tend to be more satisfied, whereas
dissatisfaction is concentrated among older, middle-aged demographics. Gender also plays a
role, as female customers show a higher engagement and a slightly higher mean satisfaction
score (3.235) compared to males (3.125). Finally, education level significantly influences
feedback complexity: University graduates demonstrate a polarized response (splitting
between high satisfaction and dissatisfaction), while College graduates appear to be the most
critical group, clustering around neutral or negative ratings.

4.2.4. Relationship between Satisfaction and Demographic factors of fan

integrating room deodorizing function (Product 3)

[Link]. Usage frequency

To examine the relationship between the satisfaction of users with the usage frequency, we
use an Excel function to compute the correlation coefficient and the scatter diagram to have a
visualized view of the trend.
a) Correlation coefficient Excel function
The correlation coefficient turns out to be rs= 0,056760
b) Scatter diagram
Here is the scatter diagram with the vertical axis appearing as the levels of satisfaction and
the horizontal axis representing the frequency of Product 3 being used.
 22

Figure [Link]. Scatter Diagram of satisfaction score and usage frequency of Product 3

The Excel function and diagram illustrate an extremely weak relationship between the
Satisfaction Score and the Usage Frequency for Product 3. As evidenced by rs= 0.056760,
which approaches zero, and a widely dispersed scatter diagram, this suggests that frequent
usage is not a strong predictor for satisfaction. Therefore, customer satisfaction with
Product 3 is likely influenced more by other factors than usage frequency.
[Link]. Gender
To compare and identify the relationship between customers’ gender and satisfaction score of
Product 3, we use the mean of satisfaction and the side-by-side bar chart to visualize the
relationship.
Firstly, the mean of satisfaction score across genders is being displayed in the table below.

Gender Mean of Satisfaction score

Female 3.2647

Male 3.3125
Table [Link]. Mean of satisfaction score of each gender.

As it is shown in table [Link]., both genders show an overall positive perception (above the
neutral score of 3). But with a 0.0478 difference, it can be seen that male customers show
more general satisfaction with Product 3 than that of female. Meanwhile, the side-by-side bar
chart below shows a pattern where female users tend to have a bigger proportion of responses
across every satisfaction score. This visualization indicates that female respondents generally
show more stability and an overall positivity of satisfaction levels, whereas male respondents
express a more varied range of satisfaction.
 23

Figure [Link]. Side-by-side bar chart of gender and satisfaction score of Product 3

As a result from previous analysis, gender is not a primary factor affecting the satisfaction
score of Product 3. Although the mean satisfaction of males is slightly higher than that of
females which is only a 0.0478 difference, the gap is too small to be identified as a strong
proof of a driver factor. This finding suggests that although there are some visible variations
in distribution, the overall satisfaction levels between males and females are largely similar.

[Link]. Age
With quantitative data like age, when we compare it with satisfaction scores, we use box plot
as our main method to determine the relationship between these two variables of Product 3.
 24
Figure [Link]. Box plot represents the relationship between Age and satisfaction of Product
3

The box plot analyzes the relationship between Age and Satisfaction Score for Product 3
shows a clear demographic trend. The median age varies between groups, with Score 3
recording the lowest median at approximately 32 years old, while score 5 shows the highest
at around 42. Scores 1 and 2 share a similar median age near 40. Score 4 shows the largest
interquartile range, indicating the greatest variation in age for this group. Finally, the dataset
includes a distinct outlier in the Score 5 category, representing a single data point below 20
years old.

From this result, we indicate that youngsters tend to stay neutral about the product while
older users express both strong opinions of dissatisfaction and satisfaction score. It can be
seen that although the satisfaction is varied among ages, the middle-aged users appear to be
the most enthusiastic demographic, determining the relationship between age and
satisfaction responses.

[Link]. Education level

With Education level being categorical data, we use the same method as gender to find out
the correlation of this factor and satisfaction scores. First, we arrange both data into this table
below using the Pivot Table of Excel. In the next step, based on the arranged data, we
generate the percentage comparison table between these two factors below.

Satisfaction Score Product 3

Education Level 1 2 3 4 5
Primary school 0% 0% 0% 18.182% 0%
High School 37.5% 36.364% 20% 18.182% 40%
College 12.5% 54.545% 20% 18.182% 26.67%
University 37.5% 9.091% 0% 27.272% 20%
Postgraduate 12.5% 0% 60% 18.182% 13.33%
Grand Total 100% 100% 100% 100% 100%
Table [Link]. Percentage between Education level and Satisfaction score of product 3

It can be seen from the table that the mid-levels respondents (High school and College) have
the most varied responses across the satisfaction score while low-education level respondents
(Primary School) participate least in showing their satisfaction. The upper-levels like
 25
University and Postgraduate indicate their dominance in specific satisfaction levels with
27.272% at score 4 and 60% of score 3, respectively. However, with the maximum level of
satisfaction (Score 5), we can see that the number of higher-level evaluations are likely to
fall. This shows that the higher the education level is, the more critical the users can be.

[Link]. Income

Using the Microsoft Excel function, the correlation coefficient between income and
satisfaction score turns out to be rs= 0.154, which shows a nearly linear correlation between
income and satisfaction. This value is statistically insignificant, hinting that a user’s income
is not a strong predictor of their satisfaction with Product 3.

Figure [Link]. Scatter plot for Monthly Income and Product 3’s Satisfaction

[Link]. Conclusion
Based on the analysis of the relationship between user satisfaction, usage frequency and
demographic factors for Product 3, we can conclude:

First, Usage Frequency and monthly income do not have a strong link to customer
satisfaction. The scatter diagrams for both variables display nearly horizontal trend lines with
widely dispersed data points. This means that how often someone uses the product and their
financial status do not directly affect how happy they are with the product.

Second, Age shows a trend where older users tend to be more polarized in their feedback
compared to younger users, and Gender also affects satisfaction because female customers
 26
tend to feel more satisfied and engaged in using Product 3 than male customers. Additionally,
consumers with higher education levels give more critical responses of satisfaction.

5. CONCLUSION: DISCUSSION AND RECOMMENDATION

Based on our comparative analysis, we have identified a clear divergence in the drivers of
customer satisfaction across Te Jia’s product lines. While Product 3 leads the market and
Product 1 performs stably, Product 2 (Vacuum Robot) is the significant underperformer. A
consistent finding across all three products is that Monthly Income is not a determinant of
satisfaction; the trend lines for income remain nearly horizontal across all scatter plots (e.g.,
Figure [Link]), confirming that a customer’s financial status has no statistical relationship
with how they rate our products.

Instead, the primary drivers are split between demographics and behavior. For Product 1 and
Product 3, satisfaction is heavily influenced by the user's profile: female customers
consistently award higher ratings than males, and higher education levels
(University/Postgraduate) correlate with more critical feedback. In sharp contrast, satisfaction
for Product 2 is driven almost exclusively by Usage Frequency. Unlike the other products, the
vacuum robot shows a moderate positive correlation where increased daily usage leads to
significantly higher scores. Additionally, Product 2 suffers from a demographic mismatch,
receiving its lowest scores from younger, tech-savvy users, while older demographics show a
higher tolerance and appreciation for the device.

To optimize our market position and rescue the underperforming Product 2 and also enhance
the quality to increase customer satisfaction for Product 1 and 3, we propose the following
strategic actions:

- Pivot Product 2 to the "Silver Economy": Since our data shows that younger users
are dissatisfied while older users are more accepting, we should reposition the robot
as a "Senior Care & Hygiene Assistant." Marketing should focus on simplicity and
health benefits (insect repelling) rather than complex smart features.
- Launch a "Habit Formation" Program: Because Usage Frequency is the crucial
key driver for Product 2’s customer satisfaction, we must ensure customers use it
daily. We propose a "30-Day Onboarding" service where support staff actively help
customers set up daily cleaning schedules to force a habit change immediately after
purchase.
- Re-engage the Male Demographic with Tech-Centric Messaging: Since our
analysis for Product 1 indicates that male users report lower satisfaction compared to
females, we propose a divergent marketing strategy. While the current "wellness and
health" messaging resonates with women, we should target male segments with
technical specification-focused campaigns highlighting CADR ratings, motor
 27
efficiency, and smart-home integration capabilities to increase their perceived value
of the product.
- Simplify Interaction for the "Polarized" Older Segment: Although Product 3 is a
market leader, the data reveals that older users are polarized. This often stems from
usability frustration. We recommend refining the physical design to include larger,
tactile buttons or a "One-Touch Auto Mode," reducing the cognitive load for older
customers who may find modern interfaces too complex.

7. ACKNOWLEDGEMENT

We would like to express our genuine gratitude to our lecturer, Ms. Chu Nguyen Mong Ngoc,
for the dedicated guidance, clear instructions, and valuable knowledge provided throughout
the duration of this project. Your thoughtful feedback and constant support have helped our
group better understand the statistical concepts and apply them effectively to real-world
analysis. This report has greatly benefited from your expertise and encouragement.

Finally, we acknowledge the collective effort of every group member. Through active
participation, responsibility, and teamwork, we were able to overcome challenges and
successfully complete this report.