Retail Demand Forecast
Analyzed approaches related to data preparation, analysis, and forecasting of time series, to facilitate recommending sales & marketing strategies based on trend/seasonality effects using data from Kaggle with diverse characteristics such as stationarity, seasonality, residuals, and sales data variance. Implemented and compared results using Exponential Smoothing Forecast, Auto Regressive Integrated Moving Average, Seasonal Auto Regressive Integrated Moving Average, Facebook’s Prophet, and LSTM Neural network. Tech used: Python (pandas, scikit-learn, prophet, stats, math, TensorFlow, Keras), Exel.
Regression Analysis of Tech Pulse Index
Course: Economic Data Analysis
This project examines dynamics of High technology industries and its effect on economy of Silicon Valley, California, which includes overall employment rate, change in housing prices and consumer sentiment in geographical area of interest. Silicon Valley’s economy closely tied to tech industries growth which has effect on employment creation, more median income, better economic condition.
Additionally, this project examines technology industry trends as they relate to the recessions and explains more on 2000’s recession. Recessions had a negative impact on components of technology industries, negative growth in job creation in computer and electronic manufacturing. The tech bubble burst is a worst phase of American tech industries in early 2000s which badly affected the economy around Silicon Valley. The report examines growth in high technology and information industries and its innovations, and how it has helped in reducing overall unemployment rate, improving the overall economic conditions. Tech used: Stata, Exel.
Project score: 98/100
Vehicle Routing Problem
This was the course project 'Vehicle Routing Optimization with Capacity and Time Window constraints' for the Operations Research course at the University of Florida. I implemented optimization algorithms to model and solve supply chain vehicle routing problems and saved total cost by 17% and reduced total number of trucks purchased to 16 from 25, thus optimizing investment. Approximate cost savings were 900,000 dollars. Tech used: Python (pandas, scikit-learn, prophet, stats, math, TensorFlow, Keras), Exel.