Unlocking Data Potential
Umadevi
Welcome to My Portfolio
Embedded Files
About Me
Fueled by a deep curiosity and passion for uncovering insights, I thrive on transforming complex data into meaningful, actionable solutions. With a Master’s in Big Data Analytics and hands-on experience in data science, I bring strong expertise in statistical modeling, machine learning, natural language processing, and data visualization. My goal is to apply these skills to solve real-world problems and drive impactful business decisions.
I'm a lifelong learner who actively stays engaged with the evolving world of data science through workshops, research, and collaboration with the wider tech community. Beyond my professional interests, I find inspiration in the arts and in exploring new cultures through travel both of which broaden my perspective and fuel my creativity in approaching problems from fresh angles.
At the core of my career journey is a desire to use data not just to predict outcomes, but to create value, uncover stories, and spark innovation. Whether building predictive models, optimizing workflows, or designing NLP solutions, I’m excited to contribute to work that makes a difference.
A Glimpse into My Work
Projects
Here are some of the data science and analytics projects I’ve worked on. Each project highlights real-world problems solved using data-driven approaches, statistical methods, and machine learning techniques. Explore how I’ve applied my skills to uncover insights and build impactful solutions.
Analyzed seismic patterns across the US, focusing on hotspots like California and Alaska. Forecasted earthquake trends using ML models on historical data from 2000–2023.
Tools: Tableau, Power BI, ArcGIS, ARIMA, Exponential Smoothing, LSTM
Built a text classification model to identify defective products based on Amazon customer reviews using NLP. Improved accuracy through hyperparameter tuning and data balancing.
Tools: Python, Scikit-learn, NLP, GridSearchCV, SMOTE, ADASYN
Developed predictive models to detect potential heart diseases using patient data. Compared performance of various ML models using accuracy and AUC metrics to identify the most effective one.
Tools: R, Data Cleaning, EDA, ML Classification Models
Built a personalised music recommender using Spotify data, combining content-based and clustering methods to suggest diverse, relevant tracks and address cold-start issues.
Tools: Python, Collaborative Filtering, Content-based Filtering
Analyzed U.S. greenhouse gas emissions (1980–2019) by sector using data from the EIA. Highlighted key contributors like transportation and electricity, visualized trends with Tableau and ArcGIS, and explored state-wise impacts.
Tools: Tableau, ArcGIS, Excel, Google Sites
Analyzed COVID-19 trends in California to understand the pandemic’s impact on public health and various industries. This project focused on identifying key patterns and potential strategies to mitigate future outbreaks.
Tools: Python, Pandas, Matplotlib, Seaborn
Developed a relational database for international cricket data covering leagues, teams, players, matches, and stadiums. Created a user-friendly web interface with search forms and result displays to simulate real-world interactions.
Tools: MySQL, HTML/CSS, SQL Procedures
Built a classification model to predict client subscription to term deposits using personal, contact, and economic data helping optimise marketing by identifying likely responders.
Tools: Python, Pandas, Scikit-learn, Streamlit, PCA, SMOTE, ML Algorithms, Neural Networks
Thank you for stopping by. I'm always open to connect, so feel free to reach out anytime!
Note: This site is still under development. Please bear with any inconvenience.
Google Sites
Report abuse