Projects
Energy and Weather Data Integration Pipeline
Collaborated with 5 others to create an ETL data pipeline in Google Cloud Platform (GCP) that integrates data from five live sources of data, processing millions of rows in real-time.
Technologies:
Key Outcomes:
- Developed python code that runs on a scheduler to fetch data from APIs at regular intervals
- Created scalable pyspark code that runs on clusters to clean, process and integrate from various sources in Google Cloud Platform
- Processed millions of rows of data in real-time
An Observational Study of iPhone Users
Conducted an in-depth study of iPhone usage patterns to understand the factors that influence purchase decisions. Interviewed over 30 users and analyzed market trends to identify key drivers of adoption.
Technologies:
Key Outcomes:
- Developed a research plan and methodology
- Collaborated with a team to collect and analyze data
- Used PowerBI to visualize data and communicate findings to stakeholders
- Identified key findings, including the importance of iMessage for iPhone users
- Recommended Apple to adopt cross-platform messaging standards
An Observational Study of iPhone Users
Conducted an in-depth study of iPhone usage patterns to understand the factors that influence purchase decisions. Interviewed over 20 individuals and analyzed market trends to identify key drivers of adoption. Developed a research plan and methodology, and collaborated with a team to collect and analyze data. Used PowerBI to visualize data and communicate findings to stakeholders.
Technologies:
Key Outcomes:
- Identified key drivers of iPhone adoption
- Highlighted the importance of iMessage for iPhone users
- Recommended Apple to adopt cross-platform messaging standards
Shift Your Focus: Spreading Awareness for Safe Driving in the Digital Age
As part of the Info Challenge at UMD, collaborated with a team of 4 to analyze a survey dataset from the Washington Traffic Safety Commission (WTSC) in order to evaluate public knowledge and adherence to traffic laws enacted in 2018.
Technologies:
Key Outcomes:
- Utilized Python, Tableau, and PowerBI to identify key parameters and subgroups within the dataset
- Recommended social media campaigns on YouTube and Facebook to raise awareness among 18-29 year-olds
- Based recommendations on insights from PewResearch center social media usage dataset
Socioeconomic Disparities in California's Water Quality
Conducted a data-driven research project comparing water quality between counties in California to analyze potential socioeconomic disparities.
Technologies:
Key Outcomes:
- Utilized multiple data sources and data cleaning techniques
- Performed a two-sample t-test to assess differences in water quality parameters
- Provided insights for policy recommendations and suggested further research
TerpGymBuddy
Led the product direction for a team of 4 to design a website that recommends exercises and time slots to UMD students based on their class schedules.
Technologies:
Key Outcomes:
- Scoped out features/user requirements
- Built a product roadmap
- Created a detailed design report before implementing the project
Distance Estimation Technique for Autonomous Vehicles
Led the development of a vision-based system for autonomous vehicles to identify, track, and estimate distances from neighboring vehicles in real-time using YOLOv4 and Deep SORT algorithms.
Technologies:
Key Outcomes:
- Devised a method for estimating distance by correlating parameters of bounding boxes with the distance between vehicles
- Implemented the solution using a combination of Python, TensorFlow, and OpenCV libraries
Smart Cart
Collaborated with three others to develop a Shopping Cart that works in real-time with the Android Application.
Technologies:
Key Outcomes:
- Used Raspberry Pi, camera, LiDAR sensor as hardware and Firebase as a database
- Automated the Supermarket shopping experience for the customer using IoT
Loan Eligibility Predictor
Designed a Loan Eligibility Predictor as part of Oracle Build-A-Thon that enables banks to sanction loans easily.
Technologies:
Key Outcomes:
- Performed data pre-processing to fill missing values, removed irrelevant fields
- Used Logistic Regression Machine Learning model to predict with a high accuracy rate
Appointment Application for Hospital
Built an Android App for the Ministry of Aayush that enables patients to book an appointment with a doctor and their preferred time as part of the Smart India Hackathon in 2020.
Technologies:
Key Outcomes:
- Developed an Android application for booking doctor appointments
- Used Firebase for authentication and database