Portfolio - viraj soni | Codebasics

Powered by
Codebasics Bootcamp

viraj soni

Data Aspirant

EMAIL

[email protected]

PHONE

+91 7043518658

Download CV

About Me

Data Analytics & AI enthusiast passionate about transforming complex data into actionable business insights. As a recent BTech Computer Engineering graduate, I specialize in machine learning, deep learning, and business intelligence solutions.
With 9+ industry projects and hands-on experience from YBI Foundation and Codebasics virtual internships, I've mastered the complete data lifecycle—from wrangling raw data to deploying production-ready AI models. My expertise spans Python, SQL, Power BI, and advanced analytics, with a proven track record of building solutions that drive real business impact.
Ready to leverage data science to solve tomorrow's challenges.

1

PowerBI Project

1

SQL Project

1

Python Project

2

Machine Learning Projects

1

Deep Learning Project

Key Skills

Python

machine learning

deep learning

gen ai

powerbi

sql

My Projects

My Experience

Dec2023-Jan2024
Data Analyst-YBI Foundation
At YBI, my data analytics internship was a dynamic journey of learning and growth. From analyzing complex datasets to crafting actionable insights.

Awards & Certificate

Excel: Mother of Business Intelligence

Get Job Ready: Power BI Data Analytics for All Levels 3.0

SQL Beginner to Advanced For Data Professionals

Python: Beginner to Advanced For Data Professionals

SQL for Data Science

Math and Statistics For AI, Data Science

Master Machine Learning for Data Science & AI: Beginner to Advanced

Let's Work Together!

Feel free to get in touch with me. I am always open to discussing new projects, creative ideas or opportunities to be part of your visions.

EMAIL

[email protected]

Call Me

+91 7043518658

LinkedIn

https://www.linkedin.com/in/virajsoni8899/

GitHub

https://github.com/virajsoni8899

Deep Learning Project

Car Damage Detection

Domain/Function: Automotive

Built a deep learning model using CNN and transfer learning (ResNet18) to classify car damage as Minor, Moderate, or Severe. Achieved 90%+ accuracy and deployed a real-time prediction app using Streamlit—ideal for insurance automation and vehicle inspection use cases.

View Project
Engagement

View Project
on GitHUB

View Project
on Streamlit

Car Damage Detection

The interface allows user to upload image of damage car by uploading it or by drag and drop

Car Damage Detection

the image is uploaded and the model will predict the car is damage or not.

Project Details

In this project, I developed a deep learning model capable of classifying car damage into three categories: Minor, Moderate, and Severe. Leveraging the power of Convolutional Neural Networks (CNN) and transfer learning with ResNet18, the model achieved over 90% validation accuracy. OpenCV was used for robust image preprocessing, enabling the system to work effectively with real-world car images.

Machine Learning Project

Credit Risk Model

Domain/Function: Finance

Built an AI-powered credit scoring system with 4-tier risk classification (Poor/Average/Good/Excellent). Features real-time default probability predictions and explainable AI for business transparency. Deployed as interactive Streamlit app enabling instant loan assessments, significantly reducing ma

View Project
Engagement

View Project
on GitHUB

View Project
on

Other Tools

python

machine learning

Credit Risk Model

Engineered an intelligent credit scoring engine that instantly evaluates loan applications with 97%+ accuracy

Credit Risk Model

Built an AI-powered credit scoring system with 4-tier risk classification (Poor/Average/Good/Excellent). Features real-time default probability predic

Project Details

Developed a sophisticated AI-powered credit scoring solution featuring a 4-tier classification system (Poor, Average, Good, Excellent) similar to CIBIL scoring. The model provides real-time default probability predictions with high explainability for business insights. Built an interactive Streamlit application enabling loan officers to assess credit applications instantly using demographic and bureau data analysis. This data-driven approach significantly reduces manual review time while maintaining accuracy, laying the foundation for Straight-Through Processing (STP) implementation. The solution transforms traditional loan assessment processes through explainable AI and user-friendly interfaces.

Project video

Machine Learning Project

Health Insurance Premium Predictor

Domain/Function: Health care prediction

Built ML model achieving 97%+ accuracy for instant premium calculations based on age, BMI, smoking habits & medical history. Cloud-deployed Streamlit app enables real-time underwriter access with <10% error margin, streamlining quote processes.

View Project
Engagement

View Project
on GitHUB

View Project
on

Other Tools

python

machine learning

Health Insurance Premium Predictor

User inputs personal and health details to predict health insurance cost using a Streamlit form interface.

Health Insurance Premium Predictor

The model predicts and displays a health insurance cost of 11380 based on entered information.

Project Details

Developed an advanced machine learning model achieving 97%+ accuracy for predicting insurance premiums based on customer demographics, BMI, smoking habits, and medical history. The solution features cloud deployment with a user-friendly Streamlit interface, enabling insurance underwriters to generate accurate premium estimates instantly.
Key achievements include maintaining <10% error margin on 95% of predictions and creating a complete end-to-end pipeline from data preprocessing to production deployment. The real-time prediction system streamlines the entire quote process, eliminating lengthy manual calculations and reducing processing time from hours to seconds.
This production-ready application transforms traditional insurance underwriting through automated risk assessment and data-driven pricing strategies, significantly improving operational efficiency.

Project video

PowerBI Project

Business Insight 360

Domain/Function: Business

In my Business Insight 360 project, I've developed a comprehensive and dynamic dashboard that provides a 360-degree view of business performance, empowering stakeholders with critical insights at their fingertips. This dashboard is meticulously designed to cover key business areas, ensuring a holist

View Project
Engagement

View Project
on GitHUB

Business Insight 360

A vibrant and dynamic dashboard from the Business Insight 360 project, specifically showcasing the Sales View.

Business Insight 360

This vibrant dashboard provides a comprehensive overview of a company's financial performance for the year-to-date (YTD)

Project Details

This comprehensive dashboard provides insightful views across several critical business areas:
📊 Sales View: Analyzing sales performance with key metrics, trends, and forecasts to drive revenue growth and strategy.
💼 Finance View: Offering a detailed financial analysis, including budgeting, expenditure tracking, and profitability insights to ensure sound financial management.
📈 Market View: Providing a comprehensive market analysis to understand market trends, customer behavior, and competitive landscape for strategic decision-making.
🚚 Supply Chain View: Enhancing supply chain efficiency with insights into inventory management, order fulfillment, and logistics performance.
🏢 Executive View: A high-level overview for executives, summarizing key metrics and KPIs across all business areas for informed decision-making.
Key Highlights: 🔹 Data Integration: Seamlessly integrated data from multiple sources to provide a unified view. 🔹 Interactive Visualizations: Created dynamic and int

Project video

Live Interactive Dashboard

SQL Project

Unveiling AtliQ Hardware’s Financial and Supply Chain Dynamics

Domain/Function: Business

This project combines data modeling, cleaning, and exploratory analysis to reveal stories embedded in financial activities. Explore expenses, revenue, and strategic insights!

View Project
Engagement

Unveiling AtliQ Hardware’s Financial and Supply Chain Dynamics

A SQL query in a database management tool retrieves product sales and forecast data. The result grid displays detailed records with dates, productcode

Unveiling AtliQ Hardware’s Financial and Supply Chain Dynamics

A SQL query calculates the total gross price of monthly sales, joining sales and price data tables. The result grid displays dates and calculated

Project Details

This in-depth exploration combines data modeling, data cleaning, and exploratory data analysis (EDA) to enhance decision-making processes and operational efficiency at AtliQ Hardware. Leveraging MySQL analytics, we tackle challenges posed by Excel’s file size constraints, providing actionable insights into sales, product inventory, customer behavior, and regional trends.

Python Project

Hotel Analysis

Domain/Function: Hotel

"In this project, I utilized Python and the powerful pandas library to perform a comprehensive analysis of hotel data. By cleaning, transforming, and visualizing key metrics, I uncovered insights into booking patterns.

View Project
on GitHUB

View Project
on google colab

The bar chart illustrates the distribution of bookings across different platforms. "Others" is the most used booking platform, followed by "makeyourtr

The bar chart shows the number of occurrences in different cities. Mumbai leads, followed by Hyderabad and Bangalore, with Delhi having the fewest

Project Details

Ever wondered how hotels use data to boost their success? This project dives into AtliQ Hotels' data, uncovering hidden patterns in bookings, occupancy rates, and revenue! We'll play detective with datasets like fact_bookings and dim_hotels, revealing which room types are most popular, which cities bring in the most money, and when guests are most likely to book. Get ready for data-driven insights that could transform how hotels operate!