Selected Projects
 | Research project: Time Series Analysis for Pharmacy Drug Sales Prediction by Applying Data Mining and Machine Learning algorithms [Code files] Researched time series forecasting and compared classical algorithms with machine learning algorithms. The use of machine learning approaches gave better results compared to classical time series methods. Applied machine learning algorithms to the time series data to predict pharmacy drug volume, which is helpful for inventory management and staff scheduling. Performed hyperparameter optimization to improve performance during training of the models. Deployed the analytics results into everyday decision making process. |
 | Capstone Project [Slides] [Code files] Roofstock is an online marketplace and community created exclusively for investing in leased Single-Family Rental homes that generate cash flow day one. They provide a transparent and efficient marketplace for single-family home investments. Both being able to understand what influences a listing getting offers, and being able to predict the time to reasonable offer(s) for a particular price would help Roofstock identify listings that need help boosting their listings. This project is to provide solutions that allows for Roofstock to be able to accurately predict how long it will take for a listing to get the offer(s) at particular listing price and offer price. Then Roofstock can implement tools or strategies for hosts of listings predicted to not get an offer to adjust listing price or other attributes for their listings. |
 | Exploratory Data Analysis and Text Mining of Google Play Store Apps [Code files] The Play Store apps data has enormous potential to drive app-making businesses to success. Actionable insights can be drawn for developers to work on and capture the Android market |
 | Predictive Modeling of Readmission Rate [Report][Code files] The Johns Hopkins University School of Medicine (JHU SoM) has collected an extensive set of patient data related to treatment and discharge, and have requested analysis revolving around predicting if a discharged patient will return within 30 days. Exploration of this phenomenon could reveal a variety of actionable insights, including improved hospital scheduling and treatment plans to match forecasted patient returns. |
 | Google Online Marketing Challenge for superstar.com [Slides] Develop, implement, and execute marketing strategies to increase sales. Completed Google Search Advertising and Display Advertising certifications. Designed marketing campaigns with Google Ads. Tracked and analyzed customer behavior with Google Analytics. Results: Acquired 650 new customers, engaged over 2,000 active users, and increased client’s revenue by 30% within 3 weeks. |