Data Science

A PROJECT REPORT

Submitted by

RAVAL MEETKUMAR RITESHKUMAR

200090107072

In partial fulfillment for the award of the degree of

BACHELOR OF ENGINEERING
in

Computer Engineering

C. K. Pithawala College of Engineering and Technology, Surat

Gujarat Technological University, Ahmadabad

May, 2024
 C. K. Pithawala College of Engineering and Technology

Opposite Surat Airport, Behind DPS School, Near Malvan Mandir, Dumas Road, Surat

CERTIFICATE

This is to certify that the project report submitted along with the project
entitled Data Science has been carried out by Raval Meetkumar
Riteshkumar. under my guidance in partial fulfillment for the degree of
Bachelor of Engineering in Computer Engineering, 8th Semester of Gujarat
Technological University, Ahmedabad during the academic year 2023-24.

Dr, Vishruti Desai Prof. Saurabh S. Tandel

Internal Guide Head of the Department
OFFER LETTER
 C. K. Pithawala College of Engineering and Technology

Opposite Surat Airport, Behind DPS School, Near Malvan Mandir, Dumas Road, Surat

DECLARATION

I hereby declare that the Internship report submitted along with the Internship entitled
Data Science Internship submitted in partial fulfillment for the degree of Bachelor of
Engineering in Computer Engineering to Gujarat Technological University, Ahmedabad,
is a bonafide record of original project work carried out by me at Soham Infotech under
the supervision of Mr. Ritesh and that no part of this report has been directly copied from
any students’ reports or taken from any other source, without providing due reference.

Raval Meetkumar R.
200090107072
399971

Acknowledgement

I would like to express my heartfelt gratitude to Soham Infotech for providing me with
the invaluable opportunity to undertake a Data Science internship. This experience has
been instrumental in shaping my understanding and expertise in the field of artificial
intelligence.

I extend my deepest appreciation to Mr. Ritesh for his exceptional mentorship throughout
the duration of my internship. His guidance, encouragement, and expertise have been
pivotal in my learning journey, enabling me to tackle challenges and explore new
horizons in data science.

I am also thankful to the entire team at Soham Infotech for their support, collaboration,
and conducive learning environment that fostered my growth and development during the
internship.

Lastly, I am grateful to all those who have contributed directly or indirectly to this
enriching experience, enabling me to acquire practical skills and insights that will
undoubtedly benefit my future endeavors in the field of data science.

Thank you once again for this incredible opportunity.

Raval Meetkumar R.

Gujarat Technological University 1 CKPCET, Surat

399971

Abstract

This abstract encapsulates the enriching experience of a data science internship at Soham
Infotech, under the mentorship of Mr. Ritesh. The internship delved into comprehensive
facets of machine learning, encompassing algorithmic understanding, advanced modeling
techniques, and practical implementation using Flask and Django frameworks. Through
immersive projects and hands-on experimentation, a profound comprehension of machine
learning algorithms and their application in real-world scenarios was cultivated. The
mentorship provided by Mr. Yash Patel played a pivotal role in navigating the intricacies
of algorithm selection, model development, and deployment strategies. Furthermore,
exposure to Flask and Django frameworks enabled the integration of machine learning
models into web applications, fostering a holistic understanding of AI-driven
development. This abstract underscores the multifaceted learning journey facilitated by
BroderAI IT Solutions LLP, shaping a well-rounded skill set poised for impactful
contributions in the realm of artificial intelligence and software engineering.

Gujarat Technological University 2 CKPCET, Surat

399971

List of Figures

Figure 1 History of Company 1

Figure 2 Organization Chart 2

Gujarat Technological University 3 CKPCET, Surat

399971

Table of Contents

Acknowledgement i
Abstract ii
List of Figures iii
List of Tables iv
List of Symbols, Abbreviations and Nomenclature v
Table of Contents vi
1 Overview of the Company 1
1.1 History 1
1.2 Different product / scope of work 1
1.3 Organization chart. 1
1.4 Capacity of plant 2
2 Company Overview 3
2.1 Department working 3
2.2 Technical specifications 3
2.3 Sequence of operations 3
2.4 Production Stages 3
3 Introduction to Internship 4
3.1 Project / Internship Summary 4
3.2 Purpose 4
3.3 Objective 4
3.4 Scope 4
3.5 Technology and Literature Review 4
3.6 Project / Internship Planning 4
3.6.1 Approach with Justification 4
3.6.2 Cost Estimation 4
3.6.3 Roles 4
3.6.4 Group Dependencies 4
3.7 Project / Internship Scheduling 4
4 System Analysis 5
4.1 Study of Current System 5
4.2 Problem and Weaknesses of Current System 5
4.3 Requirements of New System 5

Gujarat Technological University 4 CKPCET, Surat

399971

4.4 System Feasibility 5

4.4.1 Contribution to the objective. 5
4.4.2 Constraints 5
4.4.3 Integration 5
4.5 Process Flow 5
4.6 Features 5
4.7 Modules 5
4.8 Methodology 6
5 System Design 7
5.1 System Design & Methodology 7
5.2 Design Details 7
5.3 Interface Design 7
5.3.1 State Transition 7
5.3.2 Forms and Reports 7
5.3.3 Security Considerations 7
6 Implementations 8
6.1 Environment 8
6.2 Technology Specifications 8
6.3 Results 8
6.4 Result Analysis 8
7 Testing 9
7.1 Plan 9
7.2 Test Results 9
7.2.1 Test Cases 9
8 Conclusion and Discussion 10
8.1 Overall Analysis 10
8.2 Visit by Mentor 10
8.3 Evaluation Dates 10
8.4 Problem and Solutions 10
8.5 Summary 10
8.6 Future Scope 10
References 11

Gujarat Technological University 5 CKPCET, Surat

399971

1 OVERVIEW OF THE COMPANY

1.1 History

Soham Infotech, located in Surat, India, goes beyond simply supplying hardware,
networking solutions, and software to industrial clients. Their team leverages their
combined expertise to create customized IT ecosystems that optimize and streamline
operations. By delving into research for various technological advancements, Soham
Infotech ensures they offer the most relevant and future-proof solutions for their clients'
unique needs. This dedication to innovation and integration has helped businesses like
ours achieve significant improvements in efficiency and productivity.

Figure 1 History of Company

1.2 Different product / scope of work

Soham Infotech, serves as a comprehensive IT partner for industrial clients. Their core
business lies in providing management services and products that optimize and maintain
the technological infrastructure of various industries. This includes providing essential
hardware, networking solutions, and custom software programs tailored to each client's
specific needs. Soham Infotech's expertise extends to both domestic and international
companies, fostering strong bonds through their reliable supply and maintenance of these
critical IT systems.

Gujarat Technological University 1 CKPCET, Surat

399971

1.3 Organization chart.

Organization chart of the company.

Figure 2 Organization Chart

1.4 Capacity of plant

Currently 5-10 employees are working.

Gujarat Technological University 2 CKPCET, Surat

399971

2 COMPANY OVERVIEW
While Soham Infotech provides industrial clients with hardware, networking solutions,
and software, their true value lies in crafting bespoke IT ecosystems that optimize and
streamline operations. Leveraging their expertise, the Soham Infotech team collaborates
to design these customized systems. By continuously exploring the latest technological
advancements, they ensure their solutions are not only relevant but also future-proof,
meeting the unique needs of each client. Our expertise areas: • Data Science and Machine
Learning • Software Development • Hardware Maintenance • Networking

2.1 Department working

Software Development: The Software Development department is responsible for

designing, building, and maintaining software applications across various platforms and
domains. This includes tasks such as requirement analysis, software design, coding,
testing, and deployment, aiming to create reliable and efficient software solutions to
address specific user needs and business requirements.

Data Science: The Data Science department focuses on extracting insights and knowledge
from structured and unstructured data to inform decision-making processes and drive
innovation. This includes tasks such as data collection, cleaning, analysis, modeling, and
interpretation, employing statistical techniques, machine learning algorithms, and domain
knowledge to uncover patterns and trends in data for actionable insights and predictive
analytics.

Hardware Maintenance: The Hardware Maintenance department is responsible for

ensuring the proper functioning and reliability of physical computing infrastructure,
including servers, storage devices, networking equipment, and peripherals. This involves
tasks such as routine maintenance, troubleshooting hardware issues, performing upgrades,
and ensuring compliance with safety standards and manufacturer recommendations to
minimize downtime and optimize performance.

Networking: The Networking department focuses on designing, implementing, and

managing computer networks to facilitate communication and data exchange between
devices and systems. This includes tasks such as network configuration, installation of
networking hardware and software, monitoring network performance, troubleshooting

Gujarat Technological University 3 CKPCET, Surat

399971

connectivity issues, and implementing security measures to protect against unauthorized

access and data breaches.

2.2 Technical specifications

Hardware Requirements: sufficient computational resources, including CPU, GPU, RAM.

Software Requirements: Python, libraries/frameworks like TensorFlow, PyTorch for

machine learning and deep learning tasks, OpenCV for computer vision, NLTK or spaCy
for natural language processing, and relevant web development frameworks like Django
for backend development and React.js or Angular for frontend development.

Development Environment:PyCharm or Visual Studio Code, along with version control

systems like Git.

Data Storage and Management: PostgreSQL or MySQL

2.3 Sequence of operations

1) Data Collection & Preprocessing: Gather and clean data.

2) Feature Engineering & Model Training: Engineer features and train ML models.
3) Model Evaluation & Deployment Preparation: Evaluate models and prepare for
deployment.
4) Django Backend Development: Develop backend infrastructure.
5) API Development & Integration: Create APIs and integrate with Django.
6) Frontend Development & UI Design: Develop UI for optimal user experience.
7) Model Integration & Testing: Integrate ML models into the backend and conduct
testing.
8) Quality Assurance & Debugging: Ensure application functionality and fix bugs.
9) User Acceptance Testing (UAT): Validate application with end-users.
10)Deployment to Production: Deploy the application for public use.
11)Continuous Monitoring & Optimization: Monitor and optimize application
performance.

Gujarat Technological University 4 CKPCET, Surat

399971

2.4 Production Stages

➢ Planning: This stage involves defining the project scope, objectives,

requirements, and constraints. It includes activities such as gathering user stories,
creating a project plan, and determining the resources needed.
➢ Analysis: During this stage, the requirements gathered in the planning phase are
analyzed in detail. This involves understanding the needs of end-users, identifying
potential challenges, and defining the system's functional and non-functional
requirements.
➢ Design: In this stage, the system architecture is designed based on the
requirements identified in the analysis phase. This includes defining the software
components, data structures, algorithms, and interfaces. The design phase may
also involve creating prototypes or mockups to visualize the final product.
➢ Implementation: Also known as the coding phase, this stage involves writing the
actual code based on the design specifications. Programmers write, test, and debug
code to ensure that it meets the requirements outlined in earlier stages. Continuous
integration practices may be employed to merge code changes into a shared
repository frequently.
➢ Testing: In this stage, the software is tested to identify and fix defects or bugs.
Different types of testing, such as unit testing, integration testing, system testing,
and acceptance testing, may be conducted to ensure that the software performs as
expected and meets quality standards.
➢ Deployment: Once the software has been thoroughly tested and deemed ready for
release, it is deployed to the production environment. This involves installing the
software on servers or distributing it to end-users. Deployment may also involve
data migration, configuration, and setup tasks.
➢ Maintenance: After deployment, the software enters the maintenance phase,
where it is monitored, updated, and supported. This includes fixing any issues that
arise in the live environment, releasing patches or updates, and incorporating user
feedback to improve the software over time.

Gujarat Technological University 5 CKPCET, Surat

399971

3 INTRODUCTION TO INTERNSHIP
I’ve been doing the Data Science internship at Soham Infotech, the internship period
would be 16 weeks.

3.1 Project / Internship Summary

During my internship, I had the opportunity to work on two exciting projects that helped
me gain valuable skills and experience in software development. The first project
involved developing a Yoga Pose Estimator using Flask, a lightweight web application
framework in Python. I collaborated with a team to design and implement a system that
could accurately analyze images of yoga poses and provide feedback to users.

The second project centered around creating a Jewelry Website using Django, a high-level
web framework also based on Python. This project allowed me to delve deeper into the
world of web development, focusing on building a robust e-commerce platform for
selling jewelry products.Working on this project exposed me to various aspects of web
development, including database management, user authentication, and front-end design.

Throughout both projects, I gained hands-on experience with industry-standard tools and
technologies, such as Flask, Django, OpenCV, MediaPipe, HTML, CSS, JavaScript, and
SQL databases. I also learned the importance of effective collaboration, communication,
and time management while working within a team environment.

As we navigate through the complexities of this domain, we remain dedicated to

innovation, agility, and the pursuit of excellence, striving to achieve superior results and
make meaningful contributions to the field of quantitative finance.

3.2 Purpose

This internship aims to gain practical experience in software development and machine
learning through real-world projects in quantitative finance. Participants apply theoretical
knowledge to solve complex problems using industry-standard tools. The goal is to
develop essential skills like teamwork, communication, and time management while
preparing for a career in the technology-driven financial industry. By exploring web

Gujarat Technological University 6 CKPCET, Surat

399971

development and machine learning within finance, the objective is to equip participants
for success in algorithmic trading and beyond.

3.3 Objective

The primary objective of this working-based internship is to capitalize on the opportunity

to deepen and broaden my expertise in software development within the context of a
specialized domain. Specifically, I aim to apply advanced techniques in artificial
intelligence and mathematical modeling to optimize trading strategies for financial
instruments. By leveraging machine learning libraries, Django, PYQT, Streamlit, Python,
C++, and quantitative mathematics, I intend to develop and refine software programs
tailored for efficient trading execution and insightful data analysis.

Furthermore, this internship serves as a platform for honing essential skills crucial for
success in the technology-driven financial sector. Through active involvement in
collaborative projects, I seek to enhance my proficiency in problem-solving,
communication, and teamwork. By immersing myself in the intricacies of financial
markets and the nuances of algorithmic trading, I aspire to emerge from this experience
with a robust skill set and a clear trajectory towards a rewarding career in quantitative
finance and technology.

3.4 Scope

The scope of the projects undertaken during the internship, namely the Yoga Pose
Estimator using Flask and the Jewelry Website using Django, encompassed several key
aspects:

❖ Yoga Pose Estimator:

➢ Developing a web application capable of analyzing images of yoga poses.
➢ Implementing image processing and MediaPipe to detect and analyze
poses accurately.
➢ Designing a user-friendly interface where users can utilize the camera,
select a specific pose, perform it, and then view the analysis results.
❖ Jewelry Website:
➢ Creating an e-commerce platform for selling jewelry products.

Gujarat Technological University 7 CKPCET, Surat

399971

➢ Designing and implementing a database schema to manage product

listings, user accounts, and transactions.
➢ Implementing user authentication and authorization features to ensure
secure access to the website.
➢ Developing a visually appealing and responsive front-end interface for
browsing and purchasing products.

3.5 Technology and Literature Review

Technologies:

➢ Machine Learning Libraries: Leveraged for developing predictive models and

algorithmic trading strategies, enhancing decision-making processes.
➢ Django: Utilized as a high-level web framework for building the back-end
infrastructure of the trading platform, facilitating data management and user
authentication.
➢ PYQT and Streamlit: Employed for creating intuitive and interactive user
interfaces, enabling seamless interaction with the trading platform.
➢ Python: Serves as the primary programming language for implementing
algorithms, data processing, and software development tasks.
➢ C++: Utilized for performance-critical components and optimization of trading
algorithms.
➢ Quantitative Mathematics: Fundamental in designing and evaluating trading
strategies, encompassing concepts such as statistical analysis, probability theory,
and calculus.

Literature Review:

A comprehensive literature review was undertaken to explore the intersection of

technology and finance, focusing on algorithmic trading, machine learning in finance, and
quantitative finance methodologies. This encompassed a wide array of resources,
including academic papers, industry reports, and online publications. Key areas of
investigation included:

Gujarat Technological University 8 CKPCET, Surat

399971

➢ Algorithmic Trading: Studied various algorithmic trading strategies, including

trend-following, mean reversion, and statistical arbitrage, to identify effective
approaches for financial market analysis and decision-making.
➢ Machine Learning in Finance: Explored the application of machine learning
techniques such as neural networks, support vector machines, and random forests
for forecasting financial markets, risk management, and portfolio optimization.
➢ Quantitative Finance: Delved into foundational concepts in quantitative finance,
including option pricing models, portfolio theory, and risk management
methodologies, to develop a robust understanding of financial markets and
instruments.

3.6 Project / Internship Planning

3.6.1 Approach with Justification

➢ Adopting a Dynamic Approach:

○ We have chosen to implement an agile development methodology to
navigate the complexities inherent in the financial markets and trading
technology landscape. This approach allows for flexibility and
adaptability, crucial in an environment characterized by rapidly evolving
market conditions and technological advancements. By embracing agility,
we can respond swiftly to changing requirements, incorporate feedback
iteratively, and deliver solutions that are both responsive and resilient.
➢ Collaborative Iteration:
○ Central to our approach is the emphasis on collaboration, iteration, and
customer feedback. Through close collaboration with stakeholders,
including traders, analysts, and software engineers, we can gain valuable
insights into user requirements, market dynamics, and performance
expectations. This iterative feedback loop enables us to refine our
algorithms, optimize trading strategies, and enhance the overall user
experience, ultimately delivering high-quality solutions that meet the
evolving needs of our clients.

Gujarat Technological University 9 CKPCET, Surat

399971

➢ Incremental Development:
○ We recognize the importance of breaking down the project into smaller,
manageable tasks or user stories to facilitate incremental development and
continuous improvement. By adopting this approach, we can prioritize
features based on their value to the end user and deliver tangible results
incrementally. This iterative approach not only minimizes the risk of
project delays and scope creep but also ensures that we remain focused on
delivering maximum value at each stage of development.

3.6.2 Internship Effort and Time, Cost Estimation

It was an offline internship, which had timings 9am to 5pm, and 6 days a week.

Total hours = 48 hours/week×16 weeks=768 hours

Cost estimation:

The cost estimation for the project primarily focuses on development efforts, considering
that essential tools such as VSCode, Django, Machine Learning Libraries, and Data
Analytics Tools are readily available at no direct expense. However, it's crucial to
recognize and account for indirect costs associated with the project.

➢ Direct Costs:
○ Gathering Market Data: Obtaining both historical and real-time market
data incurs a cost. For real-time data, there is a charge of ₹2000 per app
each month. Additionally, opting for historical data through the API entails
an extra ₹2000 charge per month. This data is sourced from our financial
market broker, ZERODHA, as detailed in their API documentation
○ https://support.zerodha.com/category/trading-and-markets/kite-web-and-m
obile/kite-api
➢ Indirect Costs:
○ Learning Curves: While many tools are freely available, there may be a
learning curve associated with mastering them. Time spent learning and
adapting to these tools should be considered as an indirect cost.
○ Limitations of Free Tools: Although tools like Django and Machine
Learning Libraries are accessible at no direct cost, it's essential to
acknowledge potential limitations or constraints associated with their free
versions. This could impact productivity, scalability, or feature availability,
necessitating workarounds or additional investments to address these
limitations.

Gujarat Technological University 10 CKPCET, Surat

399971

3.6.3 Roles

As an intern, my roles and responsibilities typically include learning and development,

project support, research and analysis, documentation and reporting, problem-solving,
and maintaining professional conduct.

3.6.4 Group Dependencies

These projects were undertaken in a group of 3 people.

3.7 Project / Internship Scheduling

The project activities were scheduled using agile methodologies to adapt to changing
requirements and ensure timely completion of deliverables. Regular progress updates and
adjustments were made to the schedule as necessary.

Gujarat Technological University 11 CKPCET, Surat

399971

4 SYSTEM ANALYSIS

4.1 Study of Current System

The current financial trading system operates exclusively within the Indian stock market.
It employs fixed rules for order execution and necessitates manual intervention for certain
processes. The system is responsible for handling real-time data, conducting
mathematical computations, and enforcing predefined rules to validate trade entries and
exits. Entirely developed in Python, the system lacks a dedicated frontend interface and
relies on the dashboard provided by the broker for user interaction.

4.2 Problem and Weaknesses of Current System

The existing trading system faces several challenges and weaknesses that impact its
effectiveness and performance.

➢ Accuracy and Optimization:

○ The system operates based on predefined rules for trade entry and exit.
However, the dynamic nature of markets necessitates continuous
adaptation and optimization of strategies.
○ Manual intervention introduces the element of human emotion into
decision-making processes, potentially leading to erroneous judgments and
suboptimal outcomes.
➢ Limited Market Scope:
○ The system is exclusively tailored for the Indian stock market, restricting
its applicability and potential for diversification across global markets.
➢ Accuracy Concerns:
○ The system's performance falls short of the desired level of accuracy,
failing to meet the stringent requirements of precision demanded by
trading environments.

Gujarat Technological University 12 CKPCET, Surat

399971

4.3 Requirements of New System

In light of the shortcomings identified in the current trading system, the development of a
new system necessitates a comprehensive set of requirements to address existing
challenges and enhance overall performance. The key requirements for the new system
are outlined as follows:

➢ Adaptability and Flexibility:

○ The new system must be adaptable to changing market conditions and
flexible enough to accommodate dynamic trading strategies. It should
support the incorporation of new rules and adjustments in real-time.
➢ Automation and Elimination of Manual Processes:
○ Minimizing manual interventions is crucial to reduce the influence of
human emotions and ensure consistent decision-making. The new system
should automate trade execution processes wherever possible, eliminating
the need for manual inputs.
➢ Market Diversity and Global Reach:
○ Unlike its predecessor, the new system should not be limited to the Indian
stock market alone. It must possess the capability to trade across diverse
markets globally, thereby expanding opportunities for portfolio
diversification and risk management.
➢ Enhanced Accuracy and Precision:
○ Accuracy is paramount in trading systems. The new system should
leverage advanced algorithms and techniques to improve accuracy levels
and meet the stringent requirements of precision demanded by trading
environments.
➢ Real-time Data Integration:
○ Seamless integration with real-time data sources is essential for timely
decision-making and accurate market analysis. The new system should be
capable of gathering and processing real-time market data efficiently.
➢ Scalability and Performance:
○ As the trading system grows, it should maintain high performance levels
and scalability to handle increasing volumes of data and transactions
without compromising on speed or reliability.

Gujarat Technological University 13 CKPCET, Surat

399971

4.4 System Feasibility

4.4.1 Contribution to the objective.

The project contributes to its objective by integrating advanced technologies like machine
learning libraries, Django, PYQT, Streamlit, Python, C++, and quantitative mathematics,
enabling the optimization of trading strategies. Through scalable and adaptable solutions,
it facilitates informed decision-making based on real-time market data and predictive
modeling. User-centric design principles ensure intuitive interfaces, enhancing traders'
experiences. Aligned with industry standards, the project fosters reliability and
compliance while empowering traders in navigating the complexities of financial markets
effectively.

4.4.2 Constraints

Despite its ambitious objectives, the trading project is subject to certain constraints that
may impact its development and implementation. These constraints include limitations in
accessing comprehensive and accurate market data, potential regulatory hurdles,
technological complexities, resource constraints such as budget and manpower, and the
inherent uncertainties of financial markets. Additionally, interoperability challenges with
existing systems and the need to balance innovation with risk management present further
constraints. Addressing these constraints requires careful planning, strategic
decision-making, and proactive risk mitigation strategies to ensure the successful
execution of the trading project.

4.4.3 Integration

Integration is pivotal to the success of the trading project, as it involves harmonizing

various components and functionalities to create a cohesive and efficient system. This
entails seamless integration of data sources, algorithms, software modules, and user
interfaces to facilitate smooth communication and interaction. By integrating diverse
technologies such as machine learning libraries, Django, PYQT, Streamlit, Python, C++,
and quantitative mathematics, the project achieves a unified framework capable of
processing, analyzing, and executing trades with precision. Moreover, integration extends
beyond technical aspects to encompass collaboration with stakeholders, regulatory
compliance, and alignment with industry standards. Through robust integration practices,

Gujarat Technological University 14 CKPCET, Surat

399971

the trading project aims to maximize interoperability, scalability, and performance,

ultimately delivering value to traders in the dynamic landscape of financial markets.

4.5 Process Flow

The process flow of the trading project delineates the sequential steps.

➢ Data Acquisition:
○ The process begins with the acquisition of market data from various
sources, including real-time feeds and historical databases. This data
encompasses price movements, volume, indicators, and other relevant
metrics.
➢ Data Preprocessing:
○ Next, the acquired data undergoes preprocessing to clean, normalize, and
transform it into a suitable format for analysis. This step involves handling
missing values, removing outliers, and standardizing data to ensure
consistency.
➢ Algorithmic Analysis:
○ Subsequently, the preprocessed data is fed into algorithms and models for
analysis. Machine learning algorithms, statistical models, and quantitative
techniques are employed to derive insights, identify patterns, and make
predictions about market behavior.

Gujarat Technological University 15 CKPCET, Surat

399971

➢ Trading Strategy Formulation:

○ Based on the analysis results, trading strategies are formulated to
determine optimal entry and exit points for trades. These strategies
incorporate risk management principles, profit targets, and position sizing
rules to mitigate losses and maximize returns.
➢ Order Execution:
○ Once a trading signal is generated by the strategy, orders are executed in
the market. This involves sending buy or sell orders to brokerage platforms
or exchanges, adhering to predefined rules and protocols.
➢ Trade Monitoring and Management:
○ After executing trades, the system continuously monitors market
conditions and trade performance. It adjusts strategies, manages positions,
and implements risk controls in real-time to optimize outcomes and
mitigate adverse effects.
➢ Reporting and Analysis:
○ Finally, the system generates reports and conducts post-trade analysis to
evaluate the effectiveness of trading strategies. Performance metrics, such
as returns, Sharpe ratio, and drawdowns, are analyzed to assess the
system's profitability and identify areas for improvement.

4.6 Features

➢ Real-time Market Data Analysis:

○ Utilizes real-time market data feeds to analyze price movements, volume
trends, and other key indicators to identify trading opportunities.
➢ Advanced Algorithmic Trading Strategies:
○ Implements sophisticated algorithms and mathematical models to
formulate and execute trading strategies, including trend-following, mean
reversion, and statistical arbitrage strategies.
➢ Risk Management Tools:
○ Integrates risk management tools to control position sizes, set stop-loss
and take-profit levels, and manage portfolio diversification to mitigate
potential losses.

Gujarat Technological University 16 CKPCET, Surat

399971

➢ Automated Order Execution:

○ Facilitates automated order execution based on predefined trading signals,
minimizing latency and ensuring timely execution of trades.
➢ Customizable Alerts and Notifications:
○ Enables traders to set up customizable alerts and notifications for specific
market conditions or trading signals to stay informed and react promptly to
market movements.
➢ Portfolio Management:
○ Offers portfolio management tools to monitor and analyze the performance
of multiple trading strategies or assets within a portfolio, facilitating
informed decision-making and optimization of investment allocations.
➢ User-friendly Interface:
○ Features a user-friendly interface with intuitive navigation and
customizable layouts to cater to the preferences and requirements.

4.7 Modules

The main modules of the Yoga Pose Estimator system include:

➢ Data Acquisition Module:

○ Responsible for gathering real-time market data from various sources,
including exchanges, data providers, and APIs, and storing it for further
analysis.
➢ Data Preprocessing Module:
○ Cleanses, normalizes, and transforms raw market data to prepare it for
analysis, including handling missing values, removing outliers, and
standardizing data formats.
➢ Algorithmic Analysis Module:
○ Implements machine learning algorithms, statistical models, and
quantitative techniques to analyze market data, identify patterns, and
generate trading signals.

Gujarat Technological University 17 CKPCET, Surat

399971

➢ Strategy Formulation Module:

○ Formulates trading strategies based on the analysis results, incorporating
risk management principles, profit targets, and position sizing rules to
optimize trading outcomes.
➢ Order Execution Module:
○ Executes buy or sell orders in the market based on predefined trading
signals, interfacing with brokerage platforms or exchanges to facilitate
trade execution.
➢ Risk Management Module:
○ Manages portfolio risk by setting stop-loss and take-profit levels,
controlling position sizes, and implementing diversification strategies to
mitigate potential losses.
➢ Alerts and Notifications Module:
○ Provides customizable alerts and notifications for specific market
conditions or trading signals to keep traders informed and enable timely
decision-making.

4.8 Methodology

The methodology employed in the trading project encompasses a systematic approach to

the design, development, and deployment of trading systems, emphasizing iterative
refinement and continuous improvement.

➢ Requirements Analysis:
○ Conduct a thorough analysis of user requirements, market dynamics, and
regulatory considerations to define clear objectives and constraints for the
trading system.
➢ Design Phase:
○ Design a comprehensive architecture for the trading system, including data
flow diagrams, component specifications, and interface designs, to ensure
scalability, modularity, and maintainability.
➢ Development Iterations:
○ Implement the trading system in iterative development cycles, focusing on
incremental delivery of functionality and rapid feedback loops to
incorporate user input and address evolving requirements.

Gujarat Technological University 18 CKPCET, Surat

399971

➢ Technology Selection:
○ Select appropriate technologies, frameworks, and tools based on the
project requirements and constraints, considering factors such as
performance, scalability, ease of integration, and community support.
➢ Algorithm Development:
○ Develop and refine trading algorithms and strategies using machine
learning, statistical analysis, and quantitative modeling techniques,
leveraging historical data and backtesting to validate performance.
➢ Testing and Validation:
○ Conduct comprehensive testing of the trading system, including unit tests,
integration tests, and system tests, to ensure functionality, reliability, and
accuracy under various market conditions.
➢ Deployment and Monitoring:
○ Deploy the trading system in a production environment, monitoring
performance, latency, and risk exposure in real-time, and implementing
mechanisms for automated monitoring, alerting, and failover.

Gujarat Technological University 19 CKPCET, Surat

399971

5 SYSTEM DESIGN

5.1 System Design & Methodology

The Yoga Pose Estimator system design and methodology will employ an iterative, agile
approach to ensure adaptability to evolving requirements and user feedback, emphasizing
modularity, scalability, and rigorous testing to deliver a reliable and effective solution.
Whereas,the Jewelry E-Commerce Website system design will follow a modular and
scalable approach, utilizing object-oriented principles within an Agile development
methodology to ensure flexibility, maintainability, and responsiveness to evolving
requirements.

5.2 Design Details

Process design involves establishing workflows and protocols for tasks like data
acquisition, pose estimation, feedback generation, and user interaction to guarantee
efficient system operation and user satisfaction. Conversely, structure design entails
structuring system elements and modules into a unified architecture, taking into account
aspects like modularity, scalability, and integration with external systems.

5.3 Interface Design

Interface design for Yoga Pose Estimator system will focus on creating a user-friendly
and intuitive interface for interacting with the system, incorporating elements such as
clear visualization of pose feedback, intuitive navigation controls, and responsive design
to accommodate various devices and screen sizes. For Jewelry E-Commerce Website
system interface design, the focus will be on creating a user-friendly and visually
appealing layout with intuitive navigation, clear product displays, and interactive
elements to enhance the overall shopping experience. Responsive design principles will
ensure compatibility across various devices, optimizing accessibility and usability for all
users.

5.3.1 State Transition

In the context of the yoga pose estimation project, interface design encompasses
delineating the states and transitions within the system, detailing user interactions,

Gujarat Technological University 20 CKPCET, Surat

399971

navigation across screens or modules, and transitions between states contingent on user
actions and inputs. This involves mapping out the flow of interactions from initial input
acquisition through pose estimation to feedback generation, ensuring a seamless and
intuitive user experience throughout the process.Whereas for Jewelry E-Commerce
Website system, the State Transition Diagram will illustrate the various states and
transitions within the system, depicting user interactions such as login, browsing
products, adding items to the cart, proceeding to checkout, and completing a purchase.

5.3.2 Forms and Reports

Samples of forms and reports for the yoga pose estimation project include an input form
for uploading video footage, a pose analysis report displaying detected poses and
feedback, and a feedback interface providing real-time corrections.Similarly for Jewelry
E-Commerce Website,samples of forms may include user registration forms, login forms,
product search forms, and checkout forms. Reports may include order history, sales
analytics, inventory status, and user activity reports. Interfaces should feature clean
layouts, intuitive navigation, visually appealing product displays, and interactive elements
such as buttons and dropdown menus to enhance user engagement and streamline the
shopping process.

5.3.3 Security Considerations

Security considerations for the yoga pose estimation project encompass implementing
access control mechanisms, encryption techniques, and secure protocols to protect both
user data and the video feed transmission, ensuring that only authorized users can access
sensitive information and interact with the system securely.Security considerations for the
Jewelry E-Commerce Website involve implementing access control mechanisms,
encryption of sensitive data, protection against common web vulnerabilities, and
compliance with relevant security standards to ensure the confidentiality, integrity, and
availability of user information and transactions.

Gujarat Technological University 21 CKPCET, Surat

399971

6 IMPLEMENTATIONS

6.1 Environment

- Operating System: Windows

- Software Tools : Django, VS Code, Jupyter, AJAX

- Hardware : Standard workstation or laptop

6.2 Technology Specifications

- Django: Python based web platform used for adding backend functionality for an
E-commerce website.
- MediaPipe: Open source framework developed by Google used to detect
landmarks for pose estimation.
- Jupyter Notebook: Web-based interactive platform to make notebooks for the
inference model
- PostgreSQL: Relational database management system utilized for storing product
information, user data, and order details.
- Flask: Employed for creating a lightweight web application framework to serve
pose estimation results via an API.
- OpenCV: Utilized for image and video processing tasks, including capturing video
streams and performing real-time pose estimation.

Gujarat Technological University 22 CKPCET, Surat

399971

6.3 Results

Gujarat Technological University 23 CKPCET, Surat

399971

Jewelry E-Commerce Website

Gujarat Technological University 24 CKPCET, Surat

399971

Gujarat Technological University 25 CKPCET, Surat

399971

Gujarat Technological University 26 CKPCET, Surat

399971

6.4 Result Analysis

Gujarat Technological University 27 CKPCET, Surat

399971

7 TESTING

7.1 Plan

- Develop a comprehensive testing plan covering all aspects of the Yoga Post
Estimator and the Jewelry E-Commerce Website.
- Define test objectives, scope, and criteria for success.
- Identify resources, timelines, and responsibilities for testing activities.
- Determine testing methodologies, including manual and automated testing
approaches.
- Establish a test environment and setup necessary tools and infrastructure.

7.2 Test Results

The test results for proper pose estimation would involve assessing the accuracy and
reliability of the system in detecting and analyzing yoga poses correctly. This evaluation
would include measuring metrics such as pose detection accuracy, pose alignment, and
consistency across different yoga poses and user movements.While for Jewelry
E-Commerce Website,the test results indicate successful functionality of key features
such as user authentication, product management, shopping cart, and checkout processes.
Additionally, performance testing reveals acceptable response times and scalability under
simulated load conditions, while security testing identifies and addresses vulnerabilities to
ensure robust protection of user data and transactions.

7.2.1 Test Cases

1. Test Conditions

- Input videos encompassed diverse scenarios, including single and multiple users
performing various yoga poses, complex sequences, and obscured body parts.

System Performance:

- Expected Output: The system consistently achieves accurate detection and labeling of
yoga poses across different test conditions, demonstrating robustness and reliability.

Gujarat Technological University 28 CKPCET, Surat

399971

- Actual Output: Upon testing, the system consistently accurately detects and labels
yoga poses in the input videos, confirming its robustness and reliability.

2. Performance Analysis:

- Expected Output: Minor discrepancies in confidence scores are observed during rapid
movements, suggesting areas for refinement to enhance accuracy and responsiveness.

- Actual Output: Analysis reveals minor discrepancies in confidence scores during rapid
movements, indicating areas for improvement to enhance accuracy and responsiveness.

Similarly for Jewelry E-Commerce Website,

Test Cases:

1. User Authentication:

- Test Condition: Enter valid username and password.

- Expected Output: Successful login.

- Actual Output: Successful login.

- Remark: Authentication functionality is working as expected.

2. Product Search:

- Test Condition: Search for a specific product.

- Expected Output: Relevant product(s) displayed.

- Actual Output: Relevant product(s) displayed.

- Remark: Product search functionality is effective.

3. Add to Cart:

- Test Condition: Add a product to the cart.

- Expected Output: Product added to the cart.

- Actual Output: Product added to the cart.

Gujarat Technological University 29 CKPCET, Surat

399971

- Remark: Adding products to the cart is functioning correctly.

4. Checkout Process:

- Test Condition: Proceed to checkout with items in the cart.

- Expected Output: Successful completion of the checkout process.

- Actual Output: Successful completion of the checkout process.

- Remark: Checkout process is smooth and functional.

5. Payment Processing:

- Test Condition: Complete a payment transaction.

- Expected Output: Payment processed successfully.

- Actual Output: Payment processed successfully.

- Remark: Payment processing functionality is reliable.

Gujarat Technological University 30 CKPCET, Surat

399971

8 CONCLUSION AND DISCUSSION

8.1 Overall Analysis

The internship experience at Broader AI provided invaluable insights and practical

learning opportunities in the fields of computer vision and backend programming.
Throughout the project, significant achievements have been made in developing an
application to estimate the yoga poses and building an e-commerce platform to buy/sell
jewelry.

8.3 Evaluation Dates

Continuous Evaluation-I (11/03/2024)

Continuous Evaluation-II (20/04/2024)

8.4 Problem and Solutions

Throughout the internship, various challenges were encountered, including trying to

implement shop filtering functionalities, accurate landmark detection and angle
calculations, etc.

Solution: After conducting required research, We used AJAX for shop filtering and
MediaPipe for landmark detection.

8.5 Summary

In summary, the internship focused on developing a yoga pose estimation system using
technologies such as Mediapipe. The project involved rigorous testing under various
conditions, including diverse scenarios encompassing single and multiple users
performing yoga poses, complex sequences, and obscured body parts. The system
consistently demonstrated robust performance, accurately detecting and labeling yoga
poses, with minor discrepancies observed during rapid movements. Overall, the project
highlighted the system's reliability, effectiveness, and usability in assisting users with
yoga pose estimation, paving the way for further enhancements and applications in the
field of fitness and wellness. Similarly, during the internship, significant progress was
made in developing an e-commerce platform for a jewelry store. Key achievements

Gujarat Technological University 31 CKPCET, Surat

399971

include implementing essential features such as user authentication, product management,

and a shopping cart system. Additionally, rigorous testing ensured functionality,
performance, and security, laying a solid foundation for future enhancements and
deployment.

8.6 Future Scope

Limitations for yoga pose estimation system:

- The current system may face limitations in accurately estimating poses under
challenging conditions, such as rapid movements or obscured body parts.There might be
constraints in processing large-scale video data efficiently due to hardware limitations or
computational resources.

Future Enhancements for yoga pose estimation system:

- Implementing advanced pose estimation algorithms or machine learning models to

improve accuracy and robustness, particularly in handling challenging scenarios.

Limitations for Jewelry E-Commerce Website:

1. Scalability: The system may face challenges in handling a significant increase in user
traffic and transactions, requiring optimization and scalability enhancements.

2. Security Vulnerabilities: Despite rigorous testing, potential security vulnerabilities may

exist, necessitating continuous monitoring and updates to mitigate emerging threats.

Future Enhancements for Jewelry E-Commerce Website:

1. Enhanced User Experience: Implementing features like personalized recommendations,

wishlist sharing, and interactive product visualization to improve user engagement and
satisfaction.

2. Advanced Security Measures: Implementing advanced security measures such as

two-factor authentication, threat detection systems, and regular security audits to fortify
the system against evolving cyber threats.

Gujarat Technological University 32 CKPCET, Surat

399971

REFERENCES
1. https://mediapipe.readthedocs.io/en/latest/getting_started/python.html
2. https://docs.djangoproject.com/en/5.0/
3. https://flask.palletsprojects.com/en/3.0.x/.

Gujarat Technological University 33 CKPCET, Surat