BUSINESS ANALYSIS FINAL ASSIGNMENT
Table of Contents
1. Introduction. 3
2. Application. 3
3. Conclusion. 8
References. 9
Appendix 1. 10
Appendix 2. 12
Business
analysis is referred to as a disciplined approach that helps in introducing as
well as managing change for an organisation. It helps to convert their
weaknesses into strengths. Any company can enhance its business growth by
achieving a secure place inside a specified marketplace. Business data analysis
helps to get competitive advantages in an efficient manner. In this part of the
assignment, market basket analysis along with association rules of the given
retail dataset is going to be discussed in a significant manner. In addition to
this, support, lift as well as confidence will be discussed properly.
As per the given dataset, the developer has analysed
the data after retrieving the required data from the given retail dataset. In
order to retrieve the data from a given dataset, it is also required to import
data packages from the library function within an environment. In this case,
the developer has imported pandas as well as a NumPy package that helps in
continuing the overall procedure in a significant manner. After retrieving the
data, reading that value is needed (Radovilsky et al. 2018). That
is why the developer has used read.excel commands to read the overall data from
the retail excel file.
Figure 1: Result of data retrieved from
the retail dataset
(Source: Created by the developer)
Therefore, it is required to remove the null values
from all the columns of the dataset. Dropna function has been used by the
developed to remove null values. At the same time, the developer to remove the
beginning as well as ending space in a proper way has used the str.strip
function. This mentioned function also helps to remove the left as well as
right-side character depending on the argument (Embarak, Embarak & Karkal, 2018).
Along with this, the researcher has dropped the rows as well as columns of
invoice number, which contain c. After completion of the cleaning procedure,
the researcher has combined the items into one transaction based on per row. It
is required to reduce the overall dataset for analysing the whole data files in
a proper way. Smaller division helps to analyse the criteria of the data set in
a significant manner.
The dropna function in python is generally used for
removing the rows as well as columns with either the null values or NaN values.
This function allows returns or provides new data frames as well as its source.
Here by using the dropna axis is provided 0 and the subset is selected that is
invoice no and lastly the inplace in set true. In the next step the df
variables invoice is presented in astype in str. This will allow items to get
consolidated and each item is transacted into rows in a single transaction per
row. Here the additional code will allow the items to compare with the results.
In the above table, several products are presented these products with
different invoices.
Therefore, a new dataset has been generated that
stores only zero values. The developer has to convert these zero values into
positive integer values as per the recommendations of the descriptions. Based
on the encode units if the x value is greater than one, then it will return
one. Along with this, if the x value less than zero then it will return zero (Khadka, 2019). If
the x value returns zero that means no recommendations have been found as per
the customer's choice. If it returns one, then it helps to demonstrate similar
products to the customers. Customers can change their opinion depending on
their requirements. The developer has chosen the two baskets for storing the
requirements of the customer. One basket stores the needs of the potential
customers whereas other baskets store the recommendations of the requirements
of the customer. It helps any business to flaunt their alternative options as
well as the preference of the customers.
Figure 3: Sum value of different baskets
as per recommendations
(Source: Created by the developer)
The developer has used association rules along with
confidence, support as well as lift. In addition to this, the developer to find
the frequent options by building association rules with the requirements has
used the apriori algorithm. If anyone selects high lift values, that means it
helps to demonstrate more frequent items based on the number of transactions as
well as product combinations (Saura, Palos-Sanchez &
Grilo, 2019). Along
with this, it is required to have a high confidence level for analysing whether
the domain is in handy mode or not. In order to enhance the business growth,
the researcher has already set a lift as well as confidence level value that
helps in providing superior quality services to the customer. For example, in
this case, the developer has already set the lift value that is greater than 6
as well as confidence level 0.8.
Figure 4: Description of rules
(Source: Created by the developer)
Figure 5: Description of rules
(Source: Created by the developer)
According
to the case study, the recommendation has been made appropriately that has
shown in the below picture. For example, sporty paper cups, as well as paper
plates, are the recommendations based on Green Alarm Clocks. As per Alarm Clock
Bucket Green, the total recommendation sum is 340 whereas, for Alarm Clock Red,
the total recommendation sum is 316. Afterward, the developer has chosen
different combinations of Germany to analyse the recommendation part in a
proper way. According to this case study, the developer has chosen a lift value
that is greater than four as well as confidence level is greater than 0.5.
Based on that, developer has identified different recommendations that include Plasters
In Tin Spaceboy, Plasters In Tin Circus Parade as well as Red Trospot
Charlotte Bag.
Figure 6: Description of rules2
(Source: Created by the developer)
Figure 7: Description of rules2
(Source: Created by the developer)
As per the above discussion, the overall data analysis
has been performed based on the market basket analysis method. Along with this,
association rule has been depicted in the given code. Based on the association
rule, the developer has concluded the recommendations as per the description
that is being provided by the customer. In this way, the developer can be able
to accomplish the objective of the business as well as the customers
requirements.
Embarak, D. O., Embarak, & Karkal.
(2018). Data analysis and visualization using python. Apress.
Retrieved from: https://www.academia.edu/download/57894577/Data_Analysis_and_Visualization_Using_Python_-_Dr._Ossama_Embarak.pdf
Khadka, B. (2019). Data analysis theory and
practice: Case: Python and Excel Tools. Retrieved from: https://www.theseus.fi/bitstream/handle/10024/335764/Khadka_Birendra.pdf?sequence=2
Radovilsky, Z., Hegde, V., Acharya, A., &
Uma, U. (2018). Skills requirements of business data analytics and data science
jobs: A comparative analysis. Journal of Supply Chain and Operations
Management, 16(1), 82-101. Retrieved from:
https://www.csupom.com/uploads/1/1/4/8/114895679/v16n1p5.pdf
Saura, J. R., Palos-Sanchez, P., & Grilo, A.
(2019). Detecting indicators for startup business success: Sentiment analysis
using text data mining. Sustainability, 11(3), 917.
Retrieved from: https://www.mdpi.com/2071-1050/11/3/917/pdf
|
import
pandas as pd
from
mlxtend.frequent_patterns import apriori, association_rules
my_data=pd.read_excel('retail.xlsx')
my_data['Description']
= my_data['Description'].str.strip()
my_data.dropna(axis=0,
subset=['InvoiceNo'], inplace=True)
my_data['InvoiceNo']
= my_data['InvoiceNo'].astype('str')
my_data
= my_data[~my_data['InvoiceNo'].str.contains('C')]
basket
= (my_data[my_data['Country'] =="France"]
.groupby(['InvoiceNo', 'Description'])['Quantity']
.sum().unstack().reset_index().fillna(0)
.set_index('InvoiceNo'))
def
encode_units(x):
if x <= 0:
return 0
if x >= 1:
return 1
basket_sets
= basket.applymap(encode_units)
basket_sets.drop('POSTAGE',
inplace=True, axis=1)
frequent_itemsets
= apriori(basket_sets, min_support=0.07, use_colnames=True)
rules
= association_rules(frequent_itemsets, metric="lift",
min_threshold=1)
rules.head()
rules[
(rules['lift'] >= 6) &
(rules['confidence'] >= 0.8) ]
basket['ALARM
CLOCK BAKELIKE GREEN'].sum()
basket['ALARM
CLOCK BAKELIKE RED'].sum()
basket2
= (my_data[my_data['Country'] =="Germany"]
.groupby(['InvoiceNo', 'Description'])['Quantity']
.sum().unstack().reset_index().fillna(0)
.set_index('InvoiceNo'))
basket_sets2
= basket2.applymap(encode_units)
basket_sets2.drop('POSTAGE',
inplace=True, axis=1)
frequent_itemsets2
= apriori(basket_sets2, min_support=0.05, use_colnames=True)
rules2
= association_rules(frequent_itemsets2, metric="lift",
min_threshold=1)
rules2[
(rules2['lift'] >= 4) &
(rules2['confidence'] >= 0.5)]
|
+1 213 5100402
+61 489 995404
+44 203 7692978
4.6
4.72
4.92