Partnered with Citi through NYU's MSIS program to design a machine learning–based investment recommendation engine using Python; modeled client-bond relationships as graphs, applied PCA and KMeans clustering, and delivered a scalable system that boosted customer engagement through personalized bond suggestions.
Engineered data pipelines for preprocessing, normalization, and client similarity analysis using tools like pandas, NumPy, scikit-learn, and Matplotlib; resolved challenges related to high sparsity, cold start, and lack of customer metadata by leveraging bond characteristics and proposing continuous learning and time-aware enhancements.
Recognized for innovation in Internet-based community service by developing a data-driven tool to help NYU graduates identify optimal U.S. locations for employment by industry, addressing post-graduation job placement and student loan pressures through independent research and external data sourcing.
Forecasting Time Series Data - R
Analyzed historical data with ARIMA and ARCH models for one-step ahead forecasts to determine future financial volatility.
Created two projects on cryptocurrency (bitcoin) and stock market index (VIX) using Minitab and R language.
Projects on Tableau and Azure
Data Science for Business Analytics - Tableau | Azure ML
Defined modeling techniques to forecast mortgage delinquencies. Used Q1 mortgage origination and delinquency status dataset as data dictionary for data modeling in MS Azure. Predicted defaults among mortgages using prior data of Q1 2007.
Data Science for Business Analytics - Tableau | Azure ML
The main business question which can be addressed by mining the Telecom Churn dataset is “What are the major factors leading to customer churn and how to retain them?”
We explore customer usage patterns and determine which customers have high probability of unsubscribing from telecom provider’s services now or in future
Data Science for Business Analytics - Tableau | Azure ML
The most important question which can be answered by mining the employee attrition data set is that “What are the major factors contributing towards employee attrition at a firm”
We investigate various employee-centric and company-centric variables which play a crucial role in determining attrition rates.
High attrition rates have serious impact on business of firms as the cost of hiring new replacements is very high, employee morale goes down and there is a loss of work quality due to loss of trained/expertise employees hence affecting customer relationships.
Higher attrition also impacts the public image of a firm as a good place to work.
Understanding of the various factors leading to employee attrition, can help in designing and restructuring the company policies, incentive packages, work culture and training and motivational programs to retain employees for longer and reducing the chances of losing trained and skilled employees to competitors.
HR departments can also gather the characteristics of employees who stay with the company for a longer period and seek them while recruiting new employees.
Data Science for Business Analytics - Tableau
Examined the data across regions over time to identify various user utilization trends and further visualized using Tableau visualization libraries. Created regression models using MS Azure to find the best predictor for different features.
Data Science for Business Analytics - Tableau | Azure ML
Defined modeling techniques to forecast mortgage delinquencies. Used Q1 mortgage origination and delinquency status dataset as data dictionary for data modeling in MS Azure. Predicted defaults among mortgages using prior data of Q1 2007.
Data Science for Business Analytics - Tableau | Azure ML
Defined modeling techniques to forecast mortgage delinquencies. Used Q1 mortgage origination and delinquency status dataset as data dictionary for data modeling in MS Azure. Predicted defaults among mortgages using prior data of Q1 2007.