Data Analysis Report For An E-commerce Company

Background

Books are the perfect way to acquire education and help in improving overall vision of any human being. The readers read books in physical and virtually as well, physical being actual purchase of the book and virtual being reading it off internet using tablet, kindle, laptop etc. The books market has always been on rise in any country and hence website giant such as Amazon have the dedicated section for the books, this shows the seriousness of consumers around the Australia and their interest in Books. The online sales around the world of books have increase exponentially hence, dedicated sections enable users to order books and they are being delivered in stipulated time by the service providers.

Project Problem

The major importance of study following:

1. Profit analysis
2. Mean No’s of customer relationship with different aspects of data
3. Correlation Analysis of the data
4. Regression analysis of the data

Research Methodology

The various analysis techniques using statistical techniques and other tools provides the handful methods to conduct the data analysis on the data. To do the profit analysis we need to calculate the total monthly sales and corresponding monthly profit with respect to different attributes such as customer type, region, shipping and category of books on sale. The analysis based on description for the customer based on shipping, customer’s region and customer type. The p-test, two sample test and one-way ANOVA are being used in testing the comparison between the customer who bought the books and other factors based on the books. The analysis is being done using Python 3.6.

Analytical Findings

In this section, we carried the following

• Description Statistics
• Two sample t test
• Correlation Analysis
• Profit Analysis
• Regression Analysis
• One-way ANOVA

Profit Analysis

The table 1 below shows that the profit analysis of monthly sales in USD, monthly profit and respective profit percentages with respect to shipping, customer, region and category of the books. (Singh, 2012).

Table 1: Profit analysis according to for shipping type, customer type, region and category

Attributes	Level	Total Monthly Sale (in $)	Total Monthly Profit (in $)	Profit Percentage
Shipping Type	FREE	20089	12210	18.84%
PAID	44734	25870	39.91%
Customer Type	Existing	19124	7065	6.88%
New	83535	31015	30.21%
Region	SA	61420	22736	22.15%
WA	41239	15344	14.95%
Category	Comics & Graphic Novels	21032	7719	7.52%
Literature & Fiction	44409	16631	16.20%
Mystery, Thriller & Suspense	12239	4613	4.49%
Romance	24979	9117	8.88%

The number of conclusions that can be made from Table 1 is that the average profit earning 17% hence every book is approximately giving 17% profit to the company. With the variance in profit being calculated for different attributes of the company, the profit share by the paid shipping is higher as compared to the free shipping in the current data analysis. The new type of customer provides very high profit as compared to the existing customers, the users from South Australia provide more orders and profit as compared to the Western Australia. Among the categories the Literature and Fiction sales have maximum profit as compared to other books category types. (Weimao Ke,2015)

Descriptive statistics on No’s of Customers

With the number of customers on board, the profit is analyzed based on existing and new customers on the website, the analysis is made with respect to customers who bought the book from website with respect to shipping type, customer type and region of the customer. The analysis is done using the mean and standard deviation analysis. The total monthly sales and profit are in propositional with each other hence, we are making the analysis in the table 2 below:

Table 2: Summary statistics for numbers of customer who bought the books for shipping type, customer type, region and category

Attributes	Level	Size	Mean	Standard Deviation
Shipping Type	FREE	361	5.215	2.109
PAID	807	4.237	2.201
Customer Type	Existing	217	4.701	2.302
New	948	4.801	2.404
Region	SA	698	4.211	2.201
WA	470	5.401	2.309
Category	Comics & Graphic Novels	237	4.989	2.431
Literature & Fiction	505	4.201	2.210
Mystery, Thriller & Suspense	137	4.303	2.778
Romance	289	5.588	2.319

Assessment of Various Data Analytical Methods

Following observations can be made using the table, the calculation done:

1. With the average 7.6 of the mean of customer buying the books on the website.
2. The mean of customer buying, the books sales with free shipping is more than the paid shipping type.
3. Mean of new customers is slightly higher than the existing customers.
4. Mean number of customers in SA is less than WA region.
5. The romance attribute books are being preferred and mean is highest as compared to any other category.

Two Sample t-test

The analysis we are going to do is the difference between the customers who bought the book with respect to their region, shipping type and customer’s region. Let us assume that there is no significant difference among the mean of customer and different attributes basis, we would take into consideration a hypothesis there is no significant difference between the mean of the customer and different attributes under consideration and alternate hypothesis that it effects the customer buying and various other factors involved. (Morlini, 2015)

Table 3: Two sample independent test for shipping type, customer type and region

Attributes	Levels	Test Statistic	p-value
Shipping Type	Free and Paid	11.72	0.000
Customer Type	New and Existing	0.34	0.832
Region	WA and SA	-9.89	0.000

As per the table 3 we can conclude that:

1. There is lot of difference among customers who have bought the book with free and paid shipping.
2. No significant difference between the customers who were either existing or new customers.
3. We can also conclude that the difference between the customers being very high SA and WA regions.

One-way ANOVA

The customers who have bought the books in different categories would now be tested using the one-way ANOVA test, the null and alternative hypothesis have been defined below:

Null Hypothesis: there is significant difference among mean number customers who have made the purchase.

Alternative Hypothesis: there is no significant difference among mean number customers who have made the purchase

Table 4: Output of one-way ANOVA for Category

Attributes	Level	F Statistic	P Value
Category	Comics & Graphic Novels, Mystery, Thriller & Suspense, Romance and Literature & Fiction	14.12	0.000

From using the one-way ANOVA analysis test, with different categories we can make our alternative hypothesis was indeed correct and there have been different between mean number of customer and other categories from other. (Grazer, 1987)

Correlation Analysis

In this section we would be doing the correlation analysis among the various aspect of the data provided, table 5 shows the different aspects of the data analysis that have been done.

Table 5: Pearson’s correlation coefficient

	Product Price	Sale Price	Profit	Numbers of customer
Product Price	1	0.985	0.023	0.109
Sale Price	0.985	1	0.166	0.108
Profit	0.023	0.166	1	0.006
Numbers of customer	0.109	0.108	0.006	1

The following conclusions can be made using table 5:

1. The product prices are positively correlated to the profit
2. Sale price of the product is positively correlated to the profit
3. The profit is positively correlated but it has very low significance value.

Regression analysis

The regression analysis that have been done using python to predict the monthly sales of the website by the customers with books bought as our prediction variable:

Table 6: Output of Regression Analysis

F Statistic	5921.35
P Value	0.000
R2	0.832
Intercept	-12.193
Slope	38.802

As per the table 6 we can conclude that P value being 0 signifies that relationship among the monthly sales and customers who are highly significant in reaching the required target value and variables. (Morlini, Minerva & Vichi, 2015) With the value of R² being 0.832 we can suggest that the model fitting the data well and model used for the purpose is:

Total sale (in $) = -12.193 + 38.802 × No’s of Customers

Recommendations to the company

1. The mean number of customers who have made the purchases shows that the shipping of the books should be made free in order to create more orders
2. Mean number of customer is SA is lower compared to WA, the SA should be focused for the marketing process and more advertisement should be done in this region in order to increase the overall sales.
3. The mean number of Romance category of book is higher as compared to other categories than we should keep more titles of this categories in order to increase sales and overall profit.

An implementation plan based on the recommendations you have provided

1. The number of staff in logistic department should be increased in order to cater more orders
2. Using updated marketing strategies like digital marketing should be used to make the customer more aware of the current sales and products.
3. More regions should be targeted in order to sell the books
4. Titles of Romance and Literature should be expanded in order to maximize the profits.

Conclusions

The average profit earning 17% hence every book is approximately giving 17% profit to the company. With the variance in profit being calculated for different attributes of the company, the profit share by the paid shipping is higher as compared to the free shipping in the current data analysis. The new type of customer provides very high profit as compared to the existing customers, the users from South Australia provide more orders and profit as compared to the Western Australia. Among the categories the Literature and Fiction sales have maximum profit as compared to other books category types.

The average profit earning 17% hence every book is approximately giving 17% profit to the company. With the variance in profit being calculated for different attributes of the company, the profit share by the paid shipping is higher as compared to the free shipping in the current data analysis. The new type of customer provides very high profit as compared to the existing customers, the users from South Australia provide more orders and profit as compared to the Western Australia. Among the categories the Literature and Fiction sales have maximum profit as compared to other books category types.

Using the regression analysis we can conclude that P value being 0 signifies that relationship among the monthly sales and customers who are highly significant in reaching the required target value and variables. 

References

Fink, A. (2010). Advances in data analysis, data handling and business intelligence. Heidelberg: Springer.

Grazer, W., & Stiff, M. (1987). Statistical Analysis and Design in Marketing Journal Articles. Journal Of The Academy Of Marketing Science, 15(1), 70-73. doi: 10.1177/009207038701500109

Morlini, I., Minerva, T., & Vichi, M. (2015). Advances in Statistical Models for Data Analysis. Cham: Springer International Publishing.

Praveena, M., & Bharathi, B. (2017). A survey paper on big data analytics. 2017 International Conference On Information Communication And Embedded Systems (ICICES). doi: 10.1109/icices.2017.8070723

Singh, S., & Singh, N. (2012). Big Data analytics. 2012 International Conference On Communication, Information & Computing Technology (ICCICT). doi: 10.1109/iccict.2012.6398180

Weimao Ke, Borner, K., & Viswanath, L. Major Information Visualization Authors, Papers and Topics in the ACM Library. IEEE Symposium On Information Visualization. do