Statistical Analysis of Athletica Accessories Sales Data (2012-2014)

This project analyzes the sales data of Athletica Accessories, a hypothetical sports and athletic accessories manufacturer, from 2012 to 2014. The analysis utilizes a dataset collected from IBM communities, focusing on factors influencing revenue and gross profit margin across various countries and product lines. Statistical tools such as Excel and SPSS are employed to identify key predictors of revenue and profit. The project addresses three primary research questions: the significant factors impacting revenue and gross profit, the variation in gross profit margin and revenue across different countries, and whether the company maintains a gross profit margin exceeding 50%. The methodology includes regression analysis to determine influential factors, ANOVA to compare revenue and profit across countries, and t-tests to assess the company's profitability. The findings reveal significant predictors for gross margin and revenue and confirm that revenue differs by country. Additionally, the analysis supports the hypothesis that the company’s gross profit margin is greater than 50%. The project provides valuable insights for Athletica Accessories to optimize its sales strategies and improve profitability.

Running head: DATA MANAGEMENT
DATA MANAGEMENT
Name of the Student
Name of the University
Author Note

1DATA MANAGEMENT
Table of Contents
Introduction:...............................................................................................................................2
Task 1:........................................................................................................................................2
Problem definition:.................................................................................................................2
Research project planning and investigation:........................................................................2
Required Data set description:...............................................................................................3
Task 2:........................................................................................................................................3
Description of business issue:................................................................................................3
Dataset collection and preparation:........................................................................................4
Sample dataset selection, analysis and results:......................................................................4
Task 3:......................................................................................................................................12
Selection of quantitative research methodology and statistical methods:............................12
Non-effectiveness of qualitative research:...........................................................................12
Usefulness of the acquired results:.......................................................................................12
Conclusion:..............................................................................................................................13
References:...............................................................................................................................14

2DATA MANAGEMENT
Introduction:
In this business analysis project statistical analysis techniques are needed to be used for
analysing business problems of a hypothetical business. The selected business is a
manufacturer of sports and athletic accessories and the sales of the company in between the
years 2012 to 2014 are to be analysed by using suitable techniques. The company data is
collected from IBM communities and the dataset is assumed to be the sales data of a
hypothetical sports accessories company named as Athletica Accessories. The company
produces several types of accessories starting from cooking gear, tents, sleeping bags,
lanterns, rope, climbing accessories, watches and more which are listed in the collected csv
file. The company has branches in over 20 countries throughout the world and their products
are sold by various mediums including stores, web, E-mail, sales visit, telephone and other
mediums are specified in the csv file. The company also have different brand names for
similar products as sales strategy to diversify the risk of one brand’s reputations. The
products of Athletica Accessories can be classified in mainly five product lines which are
camping equipment, mountaineering equipment, personal accessories, outdoor protection and
golf equipment. The revenue and gross profit margin of the company are the main concerns
in current competitive business environment and thus it is required for the company to know
the factors which effects the profit and revenue of the company and its variation in different
regions. The statistical tools like excel and SPSS are used to analyse the sales data of the
company and suitable predictive analysis is performed to get the revenue and profit model of
the company.
Task 1:
Problem definition:
The problems of the business that are investigated in this project are
1) The significant influential factors that effects the revenue and gross profit margin of the
company.
2) Whether the gross profit margin and revenue of the company is same or significantly
different in different countries.
3) Whether the company is making more than 50% gross profit throughout the years.

3DATA MANAGEMENT
Research project planning and investigation:
In order to find answers to the research question it is required to collect a relevant data that
contains sales revenue, profit and product information of the selected company. A sales
product dataset is collected from IBM community containing a total of 88475 points having a
total of 11 variables. Among the 11 variables the variables of interest are retailer country,
order method, type of retailer, product line, product type, year, revenue, quantity and gross
margin. Now, the gross margin and revenue are considered to be resources for development
and operations of the company and hence the significant predictors for these two are required
be found using multiple regression. The regression model is assumed to be linear with
constant coefficients as no specific evidence of non-linear linear relation between the
dependent and independent variables are known.
Required Data set description:
The selected sample dataset has total 11 variables and description of each variable is given
below.
Retailer country: The countries in which the company sold their products to retailers.
Order method: Method by which orders are placed with the company.
Retailer type: Shop type of retailer
Product line: product classification
Product type: type of product on its usage
Product: name of brand of the product
Year: The year when the product is sold
Quarter: Quarter of the year when the product is sold
Revenue: Revenue in $ generated by selling the product
Quantity: The number of the products that were sold
Gross margin: The profit margin generated by selling the product (profit margin = profit/total
revenue)

4DATA MANAGEMENT
Task 2:
Description of business issue:
The issues or problems of the business of Athletica Accessories are that the factors
influencing the revenue of the business are not exactly known, only some intuitive guesses
can be formed the company director to model the revenue, however, the exact influential
factors may be different from the guesses. Also, the company needs to serve their customers
in a better way by delivering the products they need, thus it is required to found that whether
the profit margin or revenue is same or different in every countries (Kim, 2017). Moreover,
the company needs to be sure that they are earning a minimum level of profit margin by
selling their products worldwide to continue the growth of their business (Delacre, Lakens &
Leys, 2017).
Dataset collection and preparation:
Now, for finding answers to the research question problems a significant sample from the
population of entire data set of 2012-2014 is extracted for analysis. The extracted sample size
is 4500 from the population of 88000 points which satisfies minimal sampling size (as
n
N ≥ 0.05). Also, the selected sample is randomly selected from population using excel
random number generation function rand() and then sorting the entire data in ascending order
based on the random generated data from uniform distribution. Furthermore, the missing
values or incompatible values are removed by list wise basis in SPSS before performing any
statistical method. Now, for performing quantitative statistical techniques the categorical
variables are converted to specific numeric values by the ascending order of their categorical
attribute names. For example, the retailer country names from Australia to United States are
converted to numeric 1 to 21 for the 21 country names in ascending order. The similar
process is followed for all interested categorical variables.
Sample dataset selection, analysis and results:
The dataset is selected satisfying the minimum sampling size selection criterion for analysis
and all the missing or improper values are removed in list wise basis. The list wise rejection
is a procedure where the entire row of the data is removed if one or more incompleteness or
altered values are found in that row. This is very accurate and efficient as only good points
are taken for analysis and the overall data size to be analysed is reduced.
Regression analysis for finding significant factors for gross margin:

5DATA MANAGEMENT
Descriptive Statistics
Mean Std. Deviation N
Gross_margin .449840 .1196528 4475
Quantity 784.10 1547.141 4475
Country_num 11.42 6.197 4475
Retailer_num 4.90 2.317 4475
order_num 6.35 1.484 4475
product_line_num 3.35 1.729 4475
Product_type_num 10.33 6.476 4475
Revenue 40973.609111 59424.988402
5
4475
Model Summary
Model R R Square
Adjusted R
Square
Std. Error of
the Estimate
1 .211a .044 .043 .1170632
2 .211b .044 .043 .1170503
3 .210c .044 .043 .1170394
4 .210d .044 .043 .1170302
5 .210e .044 .043 .1170331
6 .209f .044 .043 .1170467
a. Predictors: (Constant), Revenue, Country_num,
Retailer_num, order_num, Product_type_num,
product_line_num, Quantity

6DATA MANAGEMENT
b. Predictors: (Constant), Revenue, Retailer_num,
order_num, Product_type_num, product_line_num, Quantity
c. Predictors: (Constant), Revenue, order_num,
Product_type_num, product_line_num, Quantity
d. Predictors: (Constant), Revenue, order_num,
Product_type_num, Quantity
e. Predictors: (Constant), Revenue, order_num, Quantity
f. Predictors: (Constant), Revenue, Quantity
Coefficientsa
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig.B Std. Error Beta
1 (Constant) .455 .010 44.699 .000
Quantity 6.103E-6 .000 .079 5.076 .000
Country_num 2.832E-5 .000 .001 .100 .920
Retailer_num .000 .001 -.006 -.409 .682
order_num .002 .001 .019 1.270 .204
product_line_num .001 .001 .008 .550 .582
Product_type_num .000 .000 -.016 -1.054 .292
Revenue -4.243E-7 .000 -.211 -13.417 .000
2 (Constant) .455 .010 47.328 .000
Quantity 6.107E-6 .000 .079 5.083 .000
Retailer_num .000 .001 -.006 -.411 .681
order_num .002 .001 .019 1.270 .204

7DATA MANAGEMENT
product_line_num .001 .001 .008 .550 .583
Product_type_num .000 .000 -.016 -1.056 .291
Revenue -4.240E-7 .000 -.211 -13.442 .000
3 (Constant) .454 .009 52.007 .000
Quantity 6.086E-6 .000 .079 5.070 .000
order_num .002 .001 .019 1.301 .193
product_line_num .001 .001 .008 .546 .585
Product_type_num .000 .000 -.016 -1.060 .289
Revenue -4.250E-7 .000 -.211 -13.513 .000
4 (Constant) .455 .008 54.969 .000
Quantity 6.068E-6 .000 .078 5.058 .000
order_num .002 .001 .021 1.400 .162
Product_type_num .000 .000 -.017 -1.106 .269
Revenue -4.273E-7 .000 -.212 -13.702 .000
5 (Constant) .452 .008 58.604 .000
Quantity 6.360E-6 .000 .082 5.434 .000
order_num .002 .001 .021 1.426 .154
Revenue -4.341E-7 .000 -.216 -14.202 .000
6 (Constant) .462 .002 209.423 .000
Quantity 6.421E-6 .000 .083 5.490 .000
Revenue -4.303E-7 .000 -.214 -14.130 .000
a. Dependent Variable: Gross_margin

8DATA MANAGEMENT
Hence, it can be seen that significant factors for gross margin by backward elimination
regression are quantity and revenue (Vu, Muttaqi & Agalgaonkar, 2015).
Regression analysis for finding significant factors for revenue:
Model Summary
Model R R Square
Adjusted R
Square
Std. Error of
the Estimate
1 .408a .166 .165 54308.076108
2
a. Predictors: (Constant), Gross_margin, order_num,
Country_num, Retailer_num, Product_type_num,
product_line_num, Quantity
Coefficientsa
Model
Unstandardized
Coefficients
Standardized
Coefficients
t Sig.B Std. Error Beta
1 (Constant) 29209.499 5665.330 5.156 .000
Quantity 10.416 .537 .271 19.388 .000
Country_num 583.509 131.277 .061 4.445 .000
Retailer_num 1743.354 351.919 .068 4.954 .000
order_num 4245.354 554.606 .106 7.655 .000
product_line_num -4047.290 477.662 -.118 -8.473 .000
Product_type_num 1553.148 128.271 .169 12.108 .000
Gross_margin -91308.201 6805.451 -.184 -13.417 .000
a. Dependent Variable: Revenue

9DATA MANAGEMENT
Hence, it is found from the backward elimination method that the significant factors for
revenue of the company are quantity, country, order type, product line type, type of product
and gross margin (Zhang & Li, 2015).
ANOVA for gross profit margin and revenue in different countries:
Test of Homogeneity of Variances
Levene
Statistic df1 df2 Sig.
Gross_margin Based on Mean .928 20 4454 .551
Based on Median .911 20 4454 .573
Based on Median and
with adjusted df
.911 20 1358.078 .573
Based on trimmed mean .896 20 4454 .593
Revenue Based on Mean 20.564 20 4479 .000
Based on Median 10.703 20 4479 .000
Based on Median and
with adjusted df
10.703 20 2255.879 .000
Based on trimmed mean 14.248 20 4479 .000
ANOVA
Sum of
Squares df Mean Square F Sig.
Gross_margin Between Groups .133 20 .007 .463 .980
Within Groups 63.921 4454 .014
Total 64.053 4474

10DATA MANAGEMENT
Revenue Between Groups 830973642740
.544
20 41548682137.
027
12.398 .000
Within Groups 150099311510
26.500
4479 3351179091.5
44
Total 158409047937
67.043
4499

11DATA MANAGEMENT
Thus it can be concluded from the significant F value 0.98 of gross margin that there is no
sufficient evidence that the gross profit margin is different in different countries and thus it is
considered that the gross profit margin is same for all the countries (Górecki, & Smaga,
2015).
However, it can be seen from the means plot of revenue and significance F value (0.00) that
the result is significant or there is sufficient evidence that the revenue is different in different
countries (Hesamian, 2016).
T-test to analyse whether Company is making more than 50% gross profit throughout
the years:
Null hypothesis (H0): The average gross profit margin of the company is less than or equal to
50% (μ<=50).
Alternative hypothesis (H1): The average gross profit margin of the company is more than
50% (μ>50) (Lakens, 2017).
One-Sample Statistics

12DATA MANAGEMENT
N Mean Std. Deviation
Std. Error
Mean
Gross_margin 4475 .449840 .1196528 .0017887
One-Sample Test
Test Value = 0.5
t df Sig. (2-tailed)
Mean
Difference
95% Confidence Interval of
the Difference
Lower Upper
Gross_margin -28.043 4474 .000 -.0501595 -.053666 -.046653
Now, the two tailed significance value is 0 or the one-tailed value is 0/2 = 0 or it can be stated
that as the p value is less than the chosen significance level of 0.05, hence there is sufficient
evidence that the gross profit margin of the company is more than 50% (Ding, Chen &
Eisenbarth, 2016).
Task 3:
Selection of quantitative research methodology and statistical methods:
The objectives of the assignment are to find the influential factors that effects the gross profit
margin and revenue of the company, finding if the average revenue and the gross profit is
same in every country and finding if the average gross profit margin is over 50%. Hence, all
of the findings require analysis with numeric data and rejecting or accepting hypothesis
statements. Thus the chosen methods are deductive quantitative methods that includes
backward elimination regression for finding significant predictors, ANOVA for comparing
means and t test for finding whether the mean gross profit is over or under a certain limit
(Brannen, 2017). Also, it is found by regression that the predictors are not explaining much of
the variation in dependent variable and hence a better dataset with better approach can
provide better results (Hepworth, 2016). Thus post-positivist approach is considered as future
scope of the research, where it is believed that the considered observation is revisable as well
as the theory and thus better results or decisions can be found in future.

13DATA MANAGEMENT
Non-effectiveness of qualitative research:
Qualitative research consists of understanding the problem and its causes through personal
opinions or based on psychological theories. As the objective of this research is to find
numeric answers or findings influential factors through numeric evidences thus qualitative
research cannot be applied in this case (McNabb, 2015). Although, the predictor variables for
revenue and gross profit margin can be predicted with intuitive opinions and thoughts but
there is no guarantee that those predictors are significant without applying quantitative
statistical methods.
Usefulness of the acquired results:
From the results obtained in task 2 it is known that the significant contributors towards gross
profit margin are quantity and revenue. Also, revenue is significantly affected by changed in
quantity, country, order type, product line type, type of product and gross product margin.
Thus the company can construct a regression model from the given coefficient table of
regression for predicting the future revenue and thus the profit margin and can focus on
specific product type, lines or even countries to sell their product in large number of
quantities. Furthermore, the ANOVA shows that the average revenue is different in different
countries but the average profit margin is approximately same in all countries. Thus the
company can country-wise select their products to be promoted or to be sold to generate high
revenue. Also, the t test suggest that company is earning more than 50% profit on an average
and thus it gives more liberty to the company for further business expansion.
Conclusion:
In this research the most obvious needs for a hypothetical business company named Athletica
Accessories that manufactures and sells sports and athletic accessories are considered as
research problems and those problems are answered with appropriate quantitative statistical
methods as discussed earlier. Now, the selected dataset only explains 16.5% of variation in
revenue and thus a better dataset can be taken to find the better model for revenue. Also,
post-hoc tests can be performed to identify for which countries revenues are different,
however, this is not performed due to scope of this report. Also, selecting a high percentage
of sample data from the population (all sales in the period of 2012 to 2014) can improve the
findings or may slightly raise the confidence level with the decisions.

14DATA MANAGEMENT
References:
Brannen, J. (2017). Mixing methods: Qualitative and quantitative research. Routledge.
Delacre, M., Lakens, D., & Leys, C. (2017). Why psychologists should by default use
Welch’s t-test instead of Student’s t-test. International Review of Social Psychology, 30(1).
Ding, A. A., Chen, C., & Eisenbarth, T. (2016, April). Simpler, faster, and more robust t-test
based leakage detection. In International workshop on constructive side-channel analysis and
secure design (pp. 163-183). Springer, Cham.
Górecki, T., & Smaga, Ł. (2015). A comparison of tests for the one-way ANOVA problem
for functional data. Computational Statistics, 30(4), 987-1010.
Hepworth, M. (2016). Research 1 Course (R1): Pre-online Handout: Glossary of key
concepts in research.
Hesamian, G. (2016). One-way ANOVA based on interval information. International Journal
of Systems Science, 47(11), 2682-2690.
Kim, T. K. (2017). Understanding one-way ANOVA using conceptual figures. Korean
journal of anesthesiology, 70(1), 22.
Lakens, D. (2017). Equivalence tests: a practical primer for t tests, correlations, and meta-
analyses. Social Psychological and Personality Science, 8(4), 355-362.
McNabb, D. E. (2015). Research methods for political science: Quantitative and qualitative
methods. Routledge.
Vu, D. H., Muttaqi, K. M., & Agalgaonkar, A. P. (2015). A variance inflation factor and
backward elimination based robust regression model for forecasting monthly electricity
demand using climatic variables. Applied Energy, 140, 385-394.
Zhang, L., & Li, K. (2015). Forward and backward least angle regression for nonlinear
system identification. Automatica, 53, 94-102.

15DATA MANAGEMENT