Rakesh Kumar Behera

Data Aspirant

[email protected] +91 8342041342

Resume

Hello, I am

Rakesh Kumar Behera.

1

PowerBI
Project

1

Python
Project

1

SQL
Project

About Me

🚀 Data Analytics Professional with 3 years of experience at Cognizant Technology Solutions and iNeuron.ai adept at transforming complex datasets into actionable business insights using advanced analytics, statistical techniques, and machine learning methodologies.

💻 Technical Expertise: Proficient in Python, SQL, Statistics, Excel, Power BI, Machine Learning, and Deep Learning, with a proven history of developing data-driven solutions that empower decision-making and deliver measurable business value.

🤖 Generative AI Enthusiast: Currently advancing expertise in Generative AI while maintaining a focus on leveraging cutting-edge technologies to tackle complex challenges and deliver innovative solutions.

💡 Strategic Communicator: Passionate about continuous learning and bridging the gap between raw data and strategic insights, with a demonstrated ability to present complex technical findings to stakeholders across all levels.

Key Skills

SQL

MS EXCEL

POWER BI

PYTHON

Machine Learning

Data Analysis

Data Visualization

Tableau

Statistics

Deep Learning

BigData

Spark

My Projects

Business 360

Domain/Function: Brick & mortar and E-Commerce

Medical Data Extraction

Domain/Function: Hospitality / Healthcare

Healthcare Insurance Premium Prediction

Domain/Function: Hospitality / Healthcare

Credit Risk Modelling

Domain/Function: Banking / Credit Risk

Apple Retail Sales Analysis

Domain/Function: Sales Analysis

Food Delivery Time Prediction

Domain/Function: Food Delivery Prediction

My Experience

Programmer Analyst | Cognizant Technology Solutions

Location: Bengaluru, India
Duration: October 2021 – May 2023

Architected and optimized SQL queries and stored procedures for streamlined reporting processes, reducing query execution time by 30% while maintaining data accuracy through regular quality checks
Automated manual data processing tasks using Python scripts and libraries, resulting in 20+ hours monthly time savings and 25% improved system efficiency through scalable data pipeline development
Enhanced team collaboration through comprehensive technical documentation while conducting regular data audits, implementing automated quality checks, and maintaining database integrity for critical business operations

Key Technologies: SQL, Python, Data Pipeline, Quality Assurance, Technical Documentation

Data Science Trainee | iNeuron.ai

Location: Bengaluru, India
Duration: July 2023 – November 2024

Acquired hands-on training in Python, SQL, Statistics, Machine Learning (NLP, Deep Learning), and AWS Cloud, designing 10+ end-to-end projects including predictive modelling, sentiment analysis (NLP), and computer vision solutions
Built and deployed scalable data pipelines and ML models (e.g., regression, neural networks) using Python libraries (Pandas, NumPy, Scikit-learn, TensorFlow) and cloud tools (AWS S3, EC2), enabling real-time insights and automation for business use cases
Applied exploratory data analysis (EDA) and Seaborn/Matplotlib visualizations to derive actionable insights, improving model accuracy by 15–25% across projects through iterative testing and optimization

Key Technologies: Python, SQL, Machine Learning, Deep Learning, NLP, AWS Cloud, Pandas, NumPy, Scikit-learn, TensorFlow, Seaborn, Matplotlib

Awards & Certificate

Excel: Mother of Business Intelligence

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

Tableau Mini

SQL Beginner to Advanced For Data Professionals

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

Python: Beginner to Advanced For Data Professionals

Full Stack Data Science Bootcamp 2.0

Let's Connect

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.

Mail Me

[email protected]

https://www.linkedin.com/in/rakesh-kumar-behera-ab2b40197/

Call Me

+91 8342041342

GitHub

https://github.com/Rakesh-Kumar-Behera

PowerBI Project

Business 360

Domain/Function: Brick & mortar and E-Commerce

Designed a multi-view dashboard in Power BI for 6 departments (Sales, Finance, Supply Chain, Executive, Marketing, and Products) of AtliQ hardware to understand sales trends, facilitate data-driven decisions that scaled the business processes by 10%.

View Project
Engagement

View Project
on GitHUB

Dashboard Home Page

Data Model

Project Details

AtliQ Hardware is a global supplier of computer peripherals. While they initially performed well, they have faced significant losses in recent years. This is primarily due to their reliance on static Excel reports, which did not provide the insights needed for informed decision-making. Consequently, they've brought in a data analytics team to develop a dynamic Power BI dashboard that serves as a multifunctional tool with insights from various departments (Finance, Sales, Marketing and Supply Chain).

𝐅𝐢𝐧𝐚𝐧𝐜𝐞 𝐕𝐢𝐞𝐰: Includes profit & loss, COGS, net sales, net profit %, gross margin.
𝐒𝐚𝐥𝐞𝐬 𝐕𝐢𝐞𝐰: Highlights gross margin % trend, product & customer performance.
𝐌𝐚𝐫𝐤𝐞𝐭𝐢𝐧𝐠 𝐕𝐢𝐞𝐰: Emphasizes gross margin % and net profit % trends, customer & product performance.
𝐒𝐮𝐩𝐩𝐥𝐲 𝐂𝐡𝐚𝐢𝐧 𝐕𝐢𝐞𝐰: Provides insights into forecast accuracy, net error, and % net error.
𝐄𝐱𝐞𝐜𝐮𝐭𝐢𝐯𝐞 𝐕𝐢𝐞𝐰: Customized for top executives with top 5 products/customers by revenu

Live Interactive Dashboard

Python Project

Medical Data Extraction

Domain/Function: Hospitality / Healthcare

An intelligent document processing system that automates the extraction of medical information from PDF prescriptions and patient records using advanced OCR techniques and computer vision.

View Project
Engagement

View Project
on GitHUB

Project Flow

Postman screenshot

Project Details

The Medical Document Data Extraction System is designed to streamline the processing of healthcare documents by converting unstructured PDF data into structured, machine-readable information. This project addresses the challenge of manually processing medical documents by automating the extraction of critical patient information, prescriptions, and medical details.

📌 Developed a robust image processing pipeline using pdf2image and OpenCV to convert and enhance medical document scans.
📌 Implemented advanced OCR capabilities using Pytesseract to accurately extract text from processed images.
📌 Built a high-performance backend API using FastAPI to handle document processing requests.
📌 Created a modular architecture that separately handles PDF conversion, image preprocessing, and text extraction.

Project video

Healthcare Insurance Premium Prediction

Domain/Function: Hospitality / Healthcare

This project focuses on predicting health insurance premiums using various regression models. The primary objective was to analyze independent features to accurately forecast premium costs. Skills: Python · Regression Analysis · Machine Learning · Scikit-Learn · Pandas · Matplotlib · Seaborn

View Project
Engagement

View Project
on GitHUB

View Project
on Streamlit

Streamlit Page to enter input

Streamlit page shows predicted premium for health insurance

Project Details

📌 Performed comprehensive exploratory data analysis using Python, NumPy, and Pandas, coupled with robust feature engineering including health risk score development and advanced encoding techniques for categorical variables.

📌 Built and optimized multiple regression models (Linear Regression, Ridge, Lasso, XGBoost), achieving initial accuracy of 92% and improving to 98% through systematic hyperparameter tuning with RandomizedSearchCV.

📌 Identified and resolved age-based prediction challenges by implementing a segmented modeling approach, creating separate models for young and older populations, reducing extreme errors from 73% to 2% in the young population segment.

📌 Achieved exceptional model performance with 99% accuracy for the older population using XGBoostRegressor, while maintaining 98% accuracy for the young population through targeted feature engineering and model refinement.

📌 The project was deployed on Streamlit.

Project video

Credit Risk Modelling

Domain/Function: Banking / Credit Risk

A supervised machine learning model using classification techniques to develop a credit risk model with a credit scorecard categorizing loan applications as Poor, Average, Good, or Excellent, akin to CIBIL scores, while predicting default probabilities and credit ratings.

View Project
Engagement

View Project
on GitHUB

View Project
on Streamlit

Credit Risk Modelling Streamlit Page for user input

Credit Risk Modelling predicted output

Project Details

📌 Exploratory Data Analysis (EDA): Conducted comprehensive EDA to ensure data integrity and uncover patterns.
📌 Data Analysis & Visualization: Leveraged Python libraries (NumPy, Pandas, Matplotlib, Seaborn) for detailed analysis and visualization.
📌 Feature Engineering: Created features like Loan-to-Income and Delinquency Ratios, applied encoding techniques, and addressed multicollinearity with VIF.
📌 Model Training: Trained Logistic Regression, RandomForestClassifier, and XGBoostClassifier, handling class imbalance with under-sampling and SMOTE-Tomek.
📌 Hyperparameter Tuning: Optimized models using RandomizedSearchCV and Optuna, achieving recall ≥90% and precision ≥50%.
📌 Model Evaluation: Evaluated models with ROC/AUC (0.98), KS Statistics, and Gini Coefficient (0.96).
🌐 Deployment on Streamlit: Delivered an interactive app for credit risk prediction.

Project video

SQL Project

Apple Retail Sales Analysis

Domain/Function: Sales Analysis

This project is designed to showcase advanced SQL querying techniques through the analysis of over 1 million rows of Apple retail sales data. The dataset includes information about products, stores, sales transactions, and warranty claims across various Apple retail locations globally.

View Project
on GitHUB

ER Diagram

Thumbnail

Project Details

This project primarily focuses on developing and showcasing the following SQL skills:

📌Complex Joins and Aggregations:
Demonstrating the ability to perform complex SQL joins and aggregate data meaningfully.
📌Window Functions:
Using advanced window functions for running totals, growth analysis, and time-based queries.
📌Data Segmentation:
Analyzing data across different time frames to gain insights into product performance.
📌Correlation Analysis:
Applying SQL functions to determine relationships between variables, such as product price and warranty claims.
📌Real-World Problem Solving:
Answering business-related questions that reflect real-world scenarios faced by data analysts.

Dataset
📌Size: 1 million+ rows of sales data.
📌Period Covered: The data spans multiple years, allowing for long-term trend analysis.
📌Geographical Coverage: Sales data from Apple stores across various countries.

Food Delivery Time Prediction

Domain/Function: Food Delivery Prediction

This project focuses on developing a food delivery time prediction model. The primary goal is to accurately estimate the time it takes for food to be delivered to customers. By providing precise delivery time predictions, food delivery platforms can enhance customer experience.

View Project
on GitHUB

Streamlit page to enter input

streaamlit page showing predicted time of food delivery

Project Details

📍1. Data Collection
- Gathered the food delivery dataset from the provided data source.
📍2. Data Preprocessing
- Performed data cleaning to handle missing values, outliers, and inconsistencies in the dataset.
- Conducted feature engineering to extract relevant features for the prediction model.
📍3. Model Development
- Utilized regression algorithms to train a food delivery time prediction model.
Explored different models such as:
- Linear Regression
- Decision Trees
- Random Forests
- XGBoost
Identified XGBoost as the best-performing model.
📍4. Model Evaluation
Evaluated model performance using the following metrics:
- Mean Squared Error (MSE)
- Root Mean Squared Error (RMSE)
- R-squared (R²) Score
📍5. Deployment
Deployed the food delivery time prediction model as a standalone Streamlit application for real-time predictions.