Data Handling: Excel, Weka, Data Mining, and Clustering Analysis

This assignment solution delves into data handling techniques using Microsoft Excel and the Weka data mining tool. Part 1 focuses on Excel, covering data pre-processing, the use of IF and LOOKUP functions, and the creation of charts and graphs for data visualization. Part 2 explores data mining, including a practical example of clustering using the audidealership.csv dataset in Weka. It explains common data mining methods like statistical techniques and decision trees, providing real-world business examples. The solution also compares the advantages and disadvantages of Weka over Excel, providing a comprehensive overview of data analysis and its applications. The solution includes the clustering model, interpretation of results, and a discussion of the strengths and weaknesses of the tools used.

DATA
HANDLING

Table of Contents
PART 1............................................................................................................................................3
1. Use of excel of pre-processing the data...................................................................................3
Use of IF function in Excel......................................................................................................4
Charts and Graphs....................................................................................................................6
PART 2............................................................................................................................................6
2.1 Using the audidealership.csv provided in conjunction with Weka give a specific example
of clustering.................................................................................................................................6
2.2 Explain the most common data mining methods that can be used in business with real
world examples..........................................................................................................................10
2.2.1 Weka J48 algorithm......................................................................................................14
2.2.2 Clustering......................................................................................................................15
2.3 Advantages and disadvantages of Weka over excel............................................................15
REFERENCES..............................................................................................................................17

PART 1
1. Use of excel of pre-processing the data
For corporate, business managers, homemakers, data collection, success reviews on a
day-to-day basis and job, companies are found in Microsoft Excel. For accounting operations
Microsoft Excel is commonly used. The following are the applications in Microsoft Excel:
 ANALYSING AND STORING DATA- Microsoft Excel is a great method for grid
research and management of data. Information may be entered in various rows and
columns of the list (Chen, Chen, He and Xia, 2018). Using maps, maps, and tables
provides data mining. The details from various archives and records can be fetched using
Excel. Data should be stored in an organized manner. This allows everyone to waste a lot
of time, other than that, it's simpler for you to analyze the results. Is not expected, using
Pivot Table, encourages data analysis (Chen and et.al., 2018).
 DATA RECOVERY- When the data is misplaced; Windows allows users to recover it.
The format MS Excel helps to restore details.
 MAKING REPORT- MS Excel and MS Word MS Word may supply the report in prose,
while MS Excel presents the report in table format. The reports are accessible in MS
Excel formats. MS Excel can allow simultaneous device distinctions. Not only is MS
Excel helping owners monitor companies, it also lets employees document their success
reviews on a regular basis. It also allows teachers use maps, patterns, methods, formulas
to assist their pupils.
 RESEARCH- They typically investigates patterns of success in the past, but success
often allows us to complete our potential work. What is the situation or how we will
continue our job this way for a few years. Research helps resolve the query about what is
going to happen and what is not to happen. Users will build prospects for our future
through different formulas and past patterns. The strategy and growth was carried out
according to this study work of corporations (Enders, 2017).
 CONDITIONAL FORMATTING- The use of conditional formatting allows to classify
all relevant people. Conditional formatting allows one to even identify issues to highlight.
It will even help to evaluate meanings, identify duplicates.

 Security- MS Excel primarily offers protections for Excel files, allowing users to manage
their data safely (Enders, 2017). Everything MS Excel files can be stored in an
outstanding repository by way of the visual basic programming. Within the MS Excel,
people store their vital data in order to maintain their data structured to save time. Most
people like to protect their files so that nobody will access them or destroy them and MS
Excel is really good at solving the issue.
Use of IF function in Excel
The IF function or IF statement in excel is made up of three parts separated by a comma.
• Condition
• When to mean when the requirements have been met
• Where to show if the conditions have not been satisfied
In this case, it can see on which point revenue and earnings have been lowered and on what date
retail purchases are that, and how the IF feature is performed step by stage.
• First copy and paste on different sheets the date of payment, revenue and income.
• Now, first rearrange old-date data by sorting excellence sheet function.
• Pick the cell where the "IF Feature" is generated.
• Type the code in the cell = if (Sort of the condition by viruses: B2>B3,)
• Enter the code which user like to see if the requirement has been reached.
• Select a comma:
• If the requirement has not been met, choose: "Increase"
• Closing the bracket and then click Enter.
• The following IF function would look like this

• When the When function displays the B3 pace or meaning than Decline and if the B3 meaning
is less than B3 than the IF function displays Raise (Sin and Gernaey, 2016). The B1 parameter
also means that the IF parameter is greater than B3.
• User should get the product of the cell B4 after you click the Enter key. Drag the handle from
D4 to cell D8400 if user wants to see the impact.
If users’ in an interview which needs strong information, user won't be shocked to first inquire
what H Lookup and V Lookup are. Whether user likes it or not, it is almost an indispensable
skill. It is quick to notice something in the details if you deal with tiny numbers. This would also
take a great deal of time to locate something in the data as the data expands (Short and Kaluvuri,
SAP SE, 2017).
Lookup Value- The base factor or search parameter in the row. That is the table's reference
point.
Table series- The table containing the values of your goals. It is raw knowledge for you to pick
from the table everything you want.
Row index number- It indicates the goal value in the row amounts. The first row is 1.
Range_ lookup]: It consists of two sets one is right (1) and the other is incorrect (0), which
looks for an identical match in the list. This is the second match in the list.
Demonstration of Look up function on given situation of Superstore Sale:
Using the same excel sheet, the following steps will be taken:
Lookup Value- The Excel LOOKUP method carries out a matching quest in a single column or
row field, and returns the equivalent value from another one column or row field (Ebtehaj,
Bonakdari and Gharabaghi, 2018). The default behavior of LOOKUP allows it easy for Excel to
fix those problems. For the date of order, revenue and income choose Cell G2, H2 and I2. On
G3, H3 and I3 the effects can be shown. Select H3 cell and place Lookup function; pick G3 cell
as Lookup key.
Table series- Choose from A2 to C8400 (A2:C8400) for the entire range.

[Range_ lookup]: Select cell for purchases, B2 to B8400 (B2:B8400).
Charts and Graphs
Steps:
• Pick a cell that is to be rendered for a line graph.
• Pick line graph to add selection.
01/01/2009
03/04/2009
04/07/2009
04/10/2009
04/01/2010
06/04/2010
07/07/2010
07/10/2010
07/01/2011
09/04/2011
10/07/2011
10/10/2011
10/01/2012
11/04/2012
12/07/2012
12/10/2012
-20000
0
20000
40000
60000
80000
100000
Sales
Profit
Interpretation: The statistic above indicates that the business has recorded the highest revenue
and lowest losses during the month of January 2009. Second highest revenues, on the other hand,
are recorded in 2012. The statistic indicates major variations in all profits and sales; in 2009 to
2010 and 2012, the business has been confronted with tremendous losses. The peak revenue
declined in 2009, although there was no downturn in other years as seen in 2009 (Sin and
Gernaey, 2016).
PART 2
2.1 Using the audidealership.csv provided in conjunction with Weka give a
specific example of clustering
=== Run information ===

Scheme: weka.clusterers.EM -I 100 -N -1 -X 10 -max -1 -ll-cv 1.0E-6 -ll-iter 1.0E-6 -M
1.0E-6 -K 10 -num-slots 1 -S 100
Relation: audidealership2
Instances: 100
Attributes: 8
Dealership
Showroom
Internet Search
RS7
A4
TT
Financing
Purchase
Test mode: evaluate on training data
=== Clustering model (full training set) ===
EM
==
Number of clusters selected by cross validation: 3
Number of iterations performed: 1
Cluster
Attribute 0 1 2
(0.34) (0.4) (0.26)
=========================================
Dealership

mean 0.4816 0.5603 0.5847
std. dev. 0.4997 0.4963 0.4928
Showroom
mean 0.9983 0.5136 0.368
std. dev. 0.0409 0.4998 0.4823
Internet Search
mean 0.3044 0.4358 0.4306
std. dev. 0.4601 0.4959 0.4952
RS7
mean 0.0008 0.886 0.6673
std. dev. 0.0281 0.3178 0.4712
A4
mean 0.9105 0.3434 0.4004
std. dev. 0.2854 0.4748 0.49
TT
mean 0.6736 0.4823 0.3003
std. dev. 0.4689 0.4997 0.4584
Financing
mean 0.5143 0.9489 0.1674
std. dev. 0.4998 0.2202 0.3733
Purchase
mean 0.2751 0.7113 0.0001
std. dev. 0.4466 0.4532 0.0084

Time taken to build model (full training data) : 0.34 seconds
=== Model and evaluation on training set ===
Clustered Instances
0 34 ( 34%)
1 28 ( 28%)
2 38 ( 38%)
Log likelihood: -2.19141

Result: The data collected above in the screenshot reveals that Audi is 1 in two, which is 52%
and 48%. Take attribute as the basis of 100; the showroom displays the highest attribute and the
lesser attribute will be shown. Depending on the study of the average and normal deviations, all
characteristics have interdependent relationships; they do not influence one another. The buying
choice of the Audi car buyer will not impact these considerations.
2.2 Explain the most common data mining methods that can be used in
business with real world examples
Data mining techniques-Data mining includes utilizing advanced data processing techniques to
identify correlations and connections in broad data sets that have been previously unknown and
true (Pashazadeh and Navimipour, 2018). These methods can involve computational simulations,
computer lectures, and mathematical algorithms like neural networks and decision-making
bodies. Data mining also requires research and estimation. Advances in Technology have

culminated in vast amounts of repositories in diverse fields. Consequently, essential information
that can be subsequently included in decision-making and company operations needs to be
processed and exploited. Data mining is the mechanism by which knowledge and trends from
enormous data are collected. The compilation, retrieval, reviews and analytics of data includes
data mining. The research or study of research / patterns is also regarded as information mining.
Information mining. In order to locate valuable data, data mining is a rational operation. If
knowledge and trends have been discovered, decisions may be made for the growth of a product.
Data mining tools will respond to your specific business concerns that have been very
challenging to overcome. We also anticipate future trends to allow traders to determine
proactively. There are three phases involved in data mining:
Exploration- During this process, the data is cleared and converted. It also decides the nature of
the results.
Pattern Recognition - The next move is to pick the pattern that allows the strongest prediction.
Deployment- To achieve the required outcomes, known trends is used.
Data mining technique- The choosing of the correct data mining technologies is one of the most
critical tasks in the industry (Blann, 2018). The technologies of data mining must be chosen
depending on the business and form of question the organization faces. To boost the precision
and cost-effectiveness of usage, a generic solution will be used. Essentially, this essay addresses
seven key data mining techniques. There are also several other forms of data mining, but these
seven are commonly used by business people.
1. Statistical Techniques- The field of statistics has been posing the same main problem as data
science for some time: how to make clear conclusions from insufficient knowledge about the
planet. A concise and reliable vocabulary to explain the relation between findings and
assumptions is among the most valuable contributions to statistics. This text carries on in the
same tradition and reflects on a variety of core problems from observational data: hypothesis
checking, trust assessment and the calculation of uncertain numbers. Data mining is a statistical
mathematics branch dealing with data collection and description. By other methods,
computational approaches are not considered by others as data mining methods. Nevertheless, it
also promotes the creation of trends and statistical models. This is why the data scientist must be

conscious of various methods in statistics. The world of today has large volumes of data to work
with and essential trends to collect. Stats can help you address questions regarding the data more
effectively.
• What is the data network pattern?
• How probable is this to occur?
• What are the most valuable trends for companies?
• What is a high level overview that offers you an insight into what the document contains?
Not only do the statistics react to these problems, they help to quantify and count the results. It
also promotes the collection of data details. There are different types of information, but the most
important and useful technique is data collection and counting. Smart decisions can be taken
through statistical reports. As other methods to gather knowledge.
2. Clustering Techniques
One of the common methods used in data mining is clustering. Analyzes of clustering is the
mechanism by which data are classified similarly (Mack, 2016). This helps to understand the
similarities and differences among the results. Often that is called segmentation which
encourages users to comprehend the activities in the database. Insurers, for example, will group
their consumers according to their employment, age, insurance structure and the form of claims.
There are different ways of clustering:
I. Division methods
II. Hierarchical agglomerative methods
III. Density based methods
IV. Grid based methods
V. Model based methods
3. View
The simplest method for identifying data trends is visualization. This approach is used at
the outset of the process of data mining. These days several academic programs are ongoing to

introduce exciting applications, known as introduce chases. There are several strategies for data
mining that build valuable trends for successful results. However, visualizing is a technology
which transforms bad data into good data in order for different types of data mining methods to
be used in the quest.
4. Induction Decision Tree Technique
A decision tree is an anticipatory pattern and its name implies it looks like a forest. Each
branch of the tree is treated in this methodology as a problem of classification, and the tree sheets
are viewed as partitions of the database connected with that specific classification. This approach
may be used for the pre-processing of data and modeling of exploratory experiments.
The decision tree can be used as a subset of the initial dataset for a specific purpose. In the
estimation of their knowledge,-data under the section reveals certain similarities. Decision trees
have outcomes which the consumer can quickly understand (Ahsan and Bais, 2018).
5. Neural Network
Neural networks are another significant methodology commonly utilized by humans.
This approach is commonly used throughout the early stages of data mining. Neural networks are
really simple to utilize as they are in fact Automated, which means that the individual is not
supposed to learn a lot about the research or software. The neural network was composed by an
artificial intelligence group. But to function effectively in neural networks, users must know:
• How connected are nodes?
• Which would be the number of production units?
• When will the exercise stop?
6. Association Rule Technique
It helps identify a link between two or more items. The interaction between different
variables in the database helps to understand. In the data collection used to classify variables and
the repeated presence of various factors with the maximum frequencies, it identifies secret
trends.

2.2.1 Weka J48 algorithm
I have 101 records in a CSV dataset and I'm using the j48 decision tree. I used the entire training
set for testing, and the results were 64 correctly classified and 37 incorrectly classified records. I
have 95 Results between percent - 100% are required and those 101 do not have the ability to
add or delete any records. But I am allowed to play a bit with my dataset, for example using 10
out of 101 in such a way that all 10 of them should be correctly classified.
 The quality of the tree is not important
 The number of test samples is not so important (10-15 is good)
 The only important thing is accuracy between 95-100%
FIRST, I tried to move the 10 correct samples to the bottom of the dataset and use "divide by
percentage" (92.30%), but it was not useful.

In seconds, I tried to pick them by "try and error" and randomly selected, and I was just testing
the last sample (101th), I got some correct samples and put it there, then the last successful
sample. Tested the last two lines of another try below, the result should have been 50% or 100%,
but was surprisingly 0% (total examples: 2 correct: 0 incorrect: 2).
2.2.2 Clustering
Interpretation: The result shows that 54% customers gives positive response and visit counter,
car and purchase; while 46% didn’t give positive response towards visiting store, counter and car
purchase.
2.3 Advantages and disadvantages of Weka over excel
The available excel sheet shows income bracket, first purchase date, last purchase date and
responded in the form of 0 and 1; where 1 implies customer look at the car and 0 implies they
didn’t look at car. The additional data or columns required to improve the data gathered are;
purchases, finance, enquiry and customer care.

The columns should be included because it will explore more approaches a customer can reach at
the stage of final purchasing of vehicle. For instance; finance and purchase will reveal how much
finance has impact on customers and purchases. Some of the excel functions like Vlookup,
Hlookup and multiple charts can be used to show the preprocessing data and interpretation of
results shown in excel charts
Weka is a programmer for open source data mining. This would not only assist with AI
estimation, it will also allow knowledge and meta-students to obtain vitality. The entire series
can be performed on any computer, written in Java. The kit comprises of three interfaces: a unix
interface, an Explorer GUI interface (which helps you to test a certain function, and calculate and
display a particular dataset and experimental GUI interface).
Advantage over excel
A package such as Weka is unquestionably desirable since a wide spectrum of details is
included, including the collection and measurement of refraction. It makes it incredibly simple to
send it a picture and display editing processes, because only one position of information is
needed. There is also an Interface package to allow its usage simpler.
Disadvantages
The most significant problem of these mining schemes is that the new techniques are not
applied. For eg, MLP performed the requisite preparation measurements and SVM does not
boost numerical estimation nor does it use the polynomial sections (Merl and Graham, 2016). It
is therefore necessary for WEKA, such as Net Lab or SVM torch, to be associated with several
different resources. The way performance is nil is a further significant drawback: the Interface
documentation is incredibly small.

REFERENCES
Books and journal:
Ahsan, U. and Bais, A., 2018. Distributed smart home architecture for data handling in smart
grid. Canadian Journal of Electrical and Computer Engineering, 41(1), pp.17-27.
Blann, A., 2018. Data handling and analysis. Oxford University Press.
Chen, C., Chen, H., He, Y. and Xia, R., 2018. TBtools, a toolkit for biologists integrating various
biological data handling tools with a user-friendly interface. BioRxiv, p.289660.
Ebtehaj, I., Bonakdari, H. and Gharabaghi, B., 2018. Development of more accurate discharge
coefficient prediction equations for rectangular side weirs using adaptive neuro-fuzzy
inference system and generalized group method of data handling. Measurement, 116,
pp.473-482.
Enders, C.K., 2017. Multiple imputation as a flexible tool for missing data handling in clinical
research. Behaviour research and therapy, 98, pp.4-18.
Mack, A., 2016. Data handling in EU-SILC.
Merl, R. and Graham, P., 2016, March. A low-cost, radiation-hardened single-board computer
for command and data handling. In 2016 IEEE Aerospace Conference (pp. 1-8). IEEE.
Pashazadeh, A. and Navimipour, N.J., 2018. Big data handling mechanisms in the healthcare
applications: A comprehensive and systematic literature review. Journal of biomedical
informatics, 82, pp.47-62.
Short, S. and Kaluvuri, S.P., SAP SE, 2017. Integrating data-handling policies into a workflow
model. U.S. Patent 9,576,257.
Sin, G. and Gernaey, K., 2016. Data handling and parameter estimation. Experimental Methods
in Wastewater Treatment, pp.201-234.