My Work
Welcome to my project page, where you can find information and updates about my ongoing initiatives. The goal is to showcase my work and provide transparency into the development process.
The Causal Effect of Education on Food Insecurity Among Adults in North Texas
Food insecurity is a significant public health concern in the U.S. Although its determinants are still being studied, the literature demonstrates that food insecurity is not related to poverty but could be correlated with financial capability. I was interested in determining if education level has a significant effect on food insecurity by using OLS in R. the data I used was collected by Dr. Millimet. The results of my analysis suggested that education does not have a statistically significant effect on food insecurity among adults and households with children under the age of 18. However, I will conduct another analysis using more advanced methods such as Probit and Logit regressions to confirm my results.
Unemployment Forecasting January 2022 - An Evaluation of Forecasting Models
As we are coming out of a global pandemic, economies are starting to recover, and we can see the U.S. Unemployment rate get to an average level. This project aims to evaluate different forecasting methods and determine which methodology is best for forecasting the unemployment rate. The methods used were Stepwise Time Series Regression and ARIMA using R. The data used was from Kaggle. It included unseasonably adjusted unemployment from 1941 to November 2021. Data from the St Louis Federal Reserve database, Fred, was used to test the accuracy of our forecast. Our experience showed that the stepwise regression was the preferred model to forecast the unemployment rate. An analysis using a Dynamic regression model will be conducted in the future to compare with our initial results.
Forecasting Potential Revenue of New Locations
The goal of the project is to help a retail client expand their business by opening new potential high-revenue locations. We had data on historical performance and the characteristics of the area around a retailer's current location. We used existing literature on determinants of sales to select variables that could significantly influence sales both statistically and economically. Using OLS regression in STATA, we found that the size of the store was the only statistical variable in our model. Distance from the highway and the competitors was economically significant but not statistically significant. After forecasting sales with the model selected, we use the MAPE (Mean Absolute Percentage Error) to determine what the actual sales of each potential store would be. the store with the highest actual sales was recommended to stakeholders during the final presentation.
Due to the confidentiality of information in the project, I will not be able to share the full document.