Key Projects

1. Comprehensive Car Sales Analysis and Visualization Using Power BI

Duration: Dec 2023 – Jan 2024

• Used Power Query to extract, transform, and load (ETL) over 250k records, optimizing data models for analysis. Applied advanced DAX functions for dynamic KPI calculations, time intelligence, and complex measures to support business requirements.
• Developed interactive dashboards with drill-through reports, cross-filtering, and custom visuals, including geospatial analysis using ArcGIS maps, allowing detailed insights into regional sales performance.
• Integrated Python scripts within Power BI for advanced data manipulation and forecasting models, enabling accurate prediction of sales trends and enhanced analytical capabilities.

2. Netflix Trend Analysis and Recommendation System

Duration: Dec 2023 – Jan 2024

• Implemented a movie recommendation system using TF-IDF vectors and cosine similarity. Developed models for predicting ratings ('Adults,' 'Kids,' 'Teens') and types ('Movies', 'TV-Show') using Naive Bayes, Logistic Regression, and Decision Tree models.
• Performed EDA on 50,000+ records with Pandas, visualizing genre trends using Matplotlib and Seaborn. Identified the top 10% of genres, leading to a 30% increase in interactions with top-rated content.
• Utilized Matplotlib and Tableau for advanced data visualizations, including dynamic dashboards and parameter controls, which facilitated interactive filtering and real-time data exploration.

3. Real-Time Weather Data ETL Automation with Apache Airflow on AWS.

Duration:July 2023 – Aug 2023

• Orchestrated a scalable ETL pipeline using Apache Airflow to process more than 50,000 weather data points daily, reducing runtime by 40% with optimized DAGs and automated task dependencies using custom Operators.
• Deployed AWS Glue for processing over 50 GB of data per day using custom PySpark jobs. Optimized data transformations, reducing processing time by 30% through dynamic partitioning and automated workflows.
• Integrated AWS Athena for querying partitioned data stored in Amazon S3, allowing for sub-second queries on 20 GB of data, improving decision-making speed by 25%, and establishing a connection of AWS Athena with AWS Quicksight to visualize weather patterns.

4. Analyzing Stock Market data in Real-Time using Kafka-driven ETL.

Duration:June 2023 – July 2023

• Developed a Python script using the yfinance library to fetch 1 year of NASDAQ stock data, iterating through ticker symbols and extracting data to flat files. Then pushed the data in S3 and launched an EC2 Linux instance for Kafka setup.
• Instantiated a Zookeeper, created a topic in Kafka, and delivered over 1,000 data messages to a consumer and S3 for ingestion into Snowflake using Kafka Connect.
• Integrated Snowflake with Snowpipe for real-time data ingestion from S3. This enabled enhanced SQL capabilities and materialized views, aiding data analysis efficiency and investment decision-making. Finally, connected Power BI to Snowflake for effective data visualization.

5. Venue Management System

Duration: Feb 2023 – May 2023

• Developed a platform for tracking user sports activities, live venue updates, and facilitating connections through a chat feature.
• Providing real-time updates using venue capacity, resulting in a 30% increase in venue utilization and a 20% boost in revenue.
• Established frontend using JavaScript and React framework, and the backend was secured on the cloud with the premium features of Firebase and then hosted on Vercel.

6. Literacy Education Nexus

Duration: Jan 2023 – April 2023

• Analyzed SOFI and World Bank data with Python and Tableau to visualize literacy education's link to global development.
• Utilized linear regression algorithms to forecast key parameters like GDP, employment, and school enrollment trends until 2040.
• Curated PowerBI visualizations to advocate education as a human right by showcasing its positive impact on economic growth.

7. Breast Cancer Detection using Machine Learning

Duration: Dec 2020 – March 2021

• For the final year capstone project, I worked on obtaining and preprocessing datasets (TCGA, Histopathology) for classifying datasets comprising images and texts.
• Under CNN, convolution and pooling reduce the layers during the pooling stage to maximize accuracy.
• Applied RFE (Recursive Feature Elimination technique) to take only the optimum features, thereby uplifting the accuracy of the model by 7-8%.

8. Fleet Management System

Duration: =Jun 2019 - Aug 2019

• Designed an online ledger using PostgreSQL to store details of heavy vehicles and drivers along with 30 days cost analysis.
• Spearheaded the analysis part for MAX Travel, leading to a 20% reduction in operational costs by visuals using Matplotlib and Seaborn.
• Not only streamlined their fleet management but also contributed to an impressive 30% reduction in vehicle downtime.

9. AWS Face Detection Application

Duration: Jun 2019 – July 2019

• Configured AWS architecture to telegram BOT. The user uploads an image to Telegram and passes the image to S3 for access.
• Actions and decisions taken are supervised by the EC2 instance, fed with an appropriate code in PHP.
• The scope was further extended in classifying gender and count of faces and other objects.

10. Blood Bank Donation System

Duration: Feb 2018 – May 2019

• Established and developed two SQL databases and schemas, one for patients’ personal history and the other for blood bank and donor details.
• Implemented comprehensive CRUD operations, guaranteeing ACID properties.
• Tested a multitude of test cases utilizing platforms such as Selenium IDE, WebDriver, and Selenium Grid, resulting in improved system performance and seamless user experience.