New Delhi,

Term Paper on “Study on Tools and methods used for Data Analysis and feature
extraction “
Submitted to – SRI RAAM COMPUTER EDUCATION
Diploma in MIS AND DATA ANALYSIS

Submitted By
Name – Dhruv Sharma

Page 1 of 15
Page 2 of 15
ACKNOWLEDGMENT
It is a high privilege for me to express my deep gratitude to those entire
faculty Members who helped me complete the project, especially my
internal guide Mr. Vipin Goswami who was always there in my hour of
need.
My special thanks to all other faculty members, Batch mates & Seniors of
SRI RAAM COMPUTER EDUCATION, Dehi for helping me in the completion
of project work and its report submission.

DHRUV SHARMA

Page 3 of 15
 TABLE OF
CONTENTS

TITLE PAGE
ACKNOWLEDGMENT

Page No.
1. Abstract………………………………………………………………………………………………………………………..4
2. Introduction………………………………………………………………………………………………………………….5
3. Data Analysis Methods………………………………………………………………………………………………………….5
4. Data Analysis Process…………………………………………………………………………………………………………..6
5. Data Analysis Techniques……………………………………………………………………………………………………..7
6. Data Analysis Tools……………………………………………………………………………………………………………….9
7. Conclusion………………………………………………………………………………………………………………….11
8. References………………………………………………………………………………………………………………….11
9. Weekly Reports…………………………………………………………………………………………………………..12

Page 4 of 15
Abstract:

This study aims to examine different tools and approaches used in data
analysis and feature extraction. The study underscores the importance of data
analytics and feature mining in these industries and provides a comprehensive
overview of the techniques regularly used in machine learning, business
analysis, and scientific research, among others. The study examined
commonly used data analysis approaches such as exploratory data analysis,
inferential statistics, regression analysis, and clustering techniques. In
addition, it analyses relevant scientific papers to support its conclusions with
case studies and concrete examples.

Page 5 of 15
Introduction
This study aims to assess the relevance of the tools and methods used in feature
extraction and data processing. It provides an overview of several exploratory
approaches to data analysis, statistical inference techniques, regression analysis
methods, clustering strategies, and commonly used feature extraction methods
such as PCA and ICA. The study also examined feature selection algorithms and
deep learning-based techniques that extract important features from large data
sets to derive relevant insights. Current research results and case studies, which
can roughly be transferred to real contexts, are used to underpin the research
results.

Data Analysis Methods

There are two basic approaches to data mining:
1. Qualitative Research: This strategy uses tools such as questionnaires,
behavioural assessments, legal implications, and other similar means to answer
the “why,” “what,” and “how” questions. Typically, these exams require written
and oral presentations, possibly including audio and video resources.
2. Quantitative Research: In this form of analysis, data is mathematically
quantified and measured. Measurement scales are used to visualize the data and
statistical adjustments are then made.
3. Text Analysis: Text analysis is a data analysis technique that aims to extract
information from text sources that is understandable to the computer. Free and
unstructured data is converted into structured data. This process involves
breaking down a collection of fragmented and unstructured documents into
manageable, understandable, and easily analyzed data segments. Text mining,
information mining, and text parsing are other names. The ambiguity of human
language is one of the difficulties of text analysis.
For example, most people know that the "Red Sox Thames Bull" has something
to do with baseball. Without prior knowledge, computers can interpret these
messages in many linguistically plausible ways. It can often be difficult for people
unfamiliar with baseball to understand the context.
4. Statistical analysis: Data collection, analysis, and verification are part of
statistics. The use of various statistical methods to quantify and analyze data is
called statistical analysis. Observational and descriptive statistics such as survey
results are examples of quantitative data. In statistical analysis, applications such
as StatSoft, SAS (Statistical Analysis System), SPSS (Statistical Package for
Social Sciences), and others are used.
5. Diagnostic Analysis: Before the statistical analysis, the diagnostic analysis
provides an in-depth investigation to solve some problems. It is also called "root

Page 6 of 15
cause analysis" and uses drill-down, drill-down, data mining, and data discovery
methods. Diagnostic analysis has three main objectives: to detect abnormalities
that require further investigation, to generate new scientific knowledge to explain
the abnormalities, and to uncover hidden connections by examining putative
causal factors.
6. Predictive Search: By applying machine learning algorithms to historical data,
predictive analytics attempts to spot important patterns and trends. Based on the
most recent data, this model is then used to predict future events. Predictive
analytics is preferred by many organizations due to its advantages such as
robustness, handling large amounts of data, good data quality, efficient and
affordable computing power, and user-friendly software. Predictive analytics is
widely used to detect fraud, optimize marketing campaigns, grow business
through wealth management and inventory forecasting, and reduce risk through
insurance approvals and creditworthiness.
7. Prescriptive Search: The prescriptive search that follows the predictive search
suggests methods and possible results that can be achieved. Because this form
of research uses standard search techniques to automatically provide
conclusions or suggestions, it requires algorithmic specificity and complex
constraints.

Data Analysis Process

Once you embark on the journey of gathering data for your study, the sheer
volume of knowledge required to make informed decisions can be overwhelming.
To arrive at well-founded conclusions and judgments amidst the abundance of
available information, it is crucial to locate the appropriate data for your study.
The following straightforward methods can assist you in finding and preparing
your data for analysis:
1. Data Requirements Specification: Define the focus of your study by
formulating a set of concise and straightforward questions that need to be
answered. Establish the criteria for measurement and describe the
variables you will be assessing as well as the intended message you wish
to convey. Set measurable goals, such as time, money, salary, etc.

2. Summary of Information: Gather data that aligns with your measurement

parameters from databases, websites, and other relevant sources. Keep in
mind that the data collected may initially be disorganized or flawed.

3. Data Processing: Once you have entered your data, add any necessary
annotations or side notes. Validate the data by comparing it with reliable
sources. Convert the data using the measurement scale established earlier

Page 7 of 15
 and eliminate any unnecessary information.

4. Data Analysis: Organize, sort, plot, and visualize correlations within your
collected data. As you proceed with organizing and processing the data,
you may discover the need to revisit certain steps, refine parameters, or
reorganize the data. Take advantage of the various data analysis tools
available to assist you in this process.

5. Discussion and Interpretation of Results: Evaluate whether the outcomes

address your original research questions. Revisit your decision-making
criteria and assess any challenges associated with implementing the
decisions. Choose an appropriate data visualization technique to
effectively communicate your message, utilizing aspects such as color
coding, layout, and graphics.

It is important to remember that any estimates or predictions derived from the

data are only educated guesses and may be subject to real-world issues.
Additionally, certain terms commonly used in data analysis can describe different
stages of the process:
1. Data Mining: This procedure involves employing techniques to identify
patterns within a sample of data.

2. Data Model: It refers to the structure and management of data within an

organization.

By employing these methods, you can navigate the process of data analysis
more effectively and draw meaningful insights from your study.

Data Analysis Techniques

There are several methods available for data mining, which can be categorized based on
the type of analysis being performed, the data being analyzed, and the volume of data
being collected. Here are some key methods in each category:

1. Numerical and Statistical-Based Methods:

- Descriptive Analysis: Analyzes performance using historical data, key
performance indicators, and predetermined criteria, taking into account current
trends and their potential impact on future outcomes.

Page 8 of 15
 - Distribution Analysis: Examines variations in factors being investigated within a
distributed database, allowing data analysts to identify variances.
- Regression Analysis: Models the relationship between a dependent variable
and one or more independent variables. It includes various regression models
such as ridge, multiple, logistic, nonlinear, survival data, etc.
- Factor Analysis: Determines relationships between variables and reveals
patterns between original variables and other factors or variables, leading to
practical grouping and classification methods.
- Differential Analysis: Categorizes data mining technique that displays various
points in different groups based on measurement variables, facilitating the
discovery of differences and promoting learning.

2. Methods based on Artificial Intelligence and Machine Learning:

- Artificial Neural Networks (ANN): Biologically inspired computer architectures
that simulate the brain's information processing. ANN can handle noisy data and
are reliable for operational categorization and forecasting purposes.
- Decision Trees: Classification or regression models represented by a tree-like
structure divide the dataset into smaller subsets based on decision rules.
- Programming using Evolutionary Algorithms: A domain-independent approach
that combines data analysis techniques, efficiently handling feature interaction
across a wide search space.
- Fuzzy Logic: A probability-based data analysis technique that deals with
uncertainty in data mining, allowing for more flexible and nuanced analysis.

Page 9 of 15
3. Methods based on Visualization and Graphs:

- Various types of charts and diagrams are utilized for data visualization, including
column charts, bar charts, line charts, pie charts, funnel charts, scatter plots,
bubble charts, Gantt charts, radar charts, and more. These visualization
techniques provide visual representations of data distributions, comparisons,
trends, and relationships, facilitating easier interpretation and understanding of
data.

These methods can be employed depending on the specific research question,

the type of data being analyzed, and the desired insights to be derived from the
analysis. By leveraging appropriate data mining techniques, researchers can
extract valuable information from their datasets and make informed decisions.

Data Analysis Tools

There are several data analytics tools available, each serving different purposes.
When selecting a tool, it is important to consider the type of study being
conducted and the nature of the data being analyzed. Here are a few useful tools
for data analysis:

1. Excel: Excel offers a wide range of features and can handle significant
amounts of data. It can be a useful tool for data analysis, especially if the data is
not too complex or close to the limits of Excel's capabilities. Learning resources
like the popular course "Data Analysis with Excel Pivot Tables" on Udemy can
help improve Excel skills.

Page 10 of 15
2. Tableau: Tableau is a powerful business intelligence tool designed specifically
for data analysis. It uses features like Pivot Tables and Pivot Charts to visually
represent data in an easily understandable manner. Tableau also provides
advanced analytics capabilities and a data cleansing tool. The online course
"Hands-On Tableau Training For Data Science" on Udemy can be a valuable
resource for learning Tableau.
3. Power BI: Originally an Excel plugin, Power BI has evolved into one of the
most robust tools for data analysis. It offers free, Pro, and Premium versions.
Power BI allows for detailed analysis, comparable to Excel formulas, using its
PowerPivot and DAX languages.
4. Fine Report: Fine Report is a reporting tool with a user-friendly drag-and-drop
interface, making it easy to create reports and develop data-driven decision
analysis processes. Its interface is similar to Excel, and it can directly connect to
various databases. Fine Report also provides a range of pre-built visual plug-in
libraries and dashboard templates.
5. R & Python: These programming languages are highly powerful and versatile.
R is commonly used for statistical analysis, including regression analysis, cluster
classification techniques, and normal distribution. It also enables personalized
predictive analytics based on browsing histories, such as customer behavior,
purchasing patterns, and product preferences. R and Python are also widely
used for artificial intelligence and machine learning.
6. SAS: SAS is a programming language specifically designed for data analysis
and manipulation. It offers various consumer-focused web, social media, and
marketing analytics products. SAS allows you to manage and optimize
interactions while making predictions about user behavior.
When selecting a data analytics tool, consider the specific requirements of your
study and choose the tool that best fits your needs in terms of data analysis
capabilities, ease of use, and compatibility with your data sources.

Page 11 of 15
Conclusion
This study explores the valuable discoveries that can be made by delving into large and
complex datasets, highlighting the importance of practical tools and effective
methodologies. It focuses on extracting actionable insights through popular data analysis
techniques, including advanced analytics for deriving meaningful knowledge from
diverse data sources. The study pays attention to market-leading technologies such as
Python, which has gained popularity and boasts strong community support, and R
statistical software, widely trusted by statisticians. It also emphasizes the usefulness of
Integrated Development Environments (IDEs) for simplifying project management tasks.

In terms of practicality and effectiveness, the study closely examines deep learning-based
contextual approaches that enhance feature extraction processes, such as ICA and PCA. It
also discusses feature selection algorithms specifically designed to achieve target
objectives while providing flexibility and empirical evidence for validating the results. To
support its findings, the study evaluates relevant research papers and real-world
applications of these tools and methods, offering valuable insights into emerging trends
and advancements over time. This research analysis is essential for researchers, data
scientists, and practitioners seeking to understand the current landscape of data analysis
techniques and their applications across different domains.

REFERENCES:
www.wikipedia.com
www.googlescholar.com

Page 12 of 15
Weekly Progress Report (WPR)

For Week 1

Program Diploma in MIS and Data Analysis Student Name Dhruv Sharma
WPR 1
Faculty Guide’s Name VIPIN GOSWAMI
Paper Title
Study on tools and methods used for Data Analysis and feature extraction.

Targets Set for the week

Gathering information about various Studies on tools and methods used for Data
Analysis and feature extraction.
Progress/Achievements for the week
Arranging the gathered information for easy accessibility.
Future Work Plans
Referring to other research papers available for more information

Weekly Progress Report (WPR)

For Week 2

Program Diploma in MIS and Data Anaylsis Student Name Dhruv Sharma
WPR 2
Faculty Guide’s Name Dr. VIPIN GOSWAMI

Paper Title
Study on tools and methods used for Data Analysis and feature extraction.

Targets Set for the week

Describing and Explaining Methods of Data Analysis

Progress/Achievements for the week

Page 13 of 15
Entering the researched data on Methods of Data Analysis

Future Work Plans

Referring to other research papers available for more information, especially on Data
Analysis Process

Page 14 of 15
Weekly Progress Report (WPR)

For Week 3

Program Diploma in MIS and Data Anaylsis Student Name Dhruv Sharma
WPR 3
Faculty Guide’s Name Dr. VIPIN GOSWAMI

Paper Title
Study on tools and methods used for Data Analysis and feature extraction.

Targets Set for the week

Describing and Explaining Data Analysis Process

Progress/Achievements for the week

Entering the researched data into Data Analysis Process
Future Work Plans
Referring to other research papers available for more information especially Data Analysis
Techniques and Data Analysis Tools

Weekly Progress Report (WPR)

For Week 4

Program Diploma in MIS and Data Anaylsis Student Name Dhruv Sharma
WPR 4
Faculty Guide’s Name Dr. VIPIN GOSWAMI

Paper Title
Study on tools and methods used for Data Analysis and feature extraction.

Targets Set for the week

Describing and explaining Data Analysis Techniques and Data Analysis Tools

Progress/Achievements for the week

Entering the researched data on Data Analysis Techniques and Data Analysis Tools

Page 15 of 15