Projects:
Below are some of the projects that I have either had the pleasure of conducting or worked on a team to complete. These projects display my ability to create data-driven insights to answer complex questions and present them in a concise, legible manner.
Please Click the Image for Full Article
Project #1
Avoiding bias when inferring race using name-based approaches
I worked on an undergraduate research team led by my research advisor, Dr. Monroe-White, to find the bias in algorithms inferring race based on people's names. In this project, I generated the data needed by collecting data and by myself, inferring based on my knowledge and life experiences, looking at the names of PHD/Masters candidates and trying to determine their race just by looking at their names. I collected data to find out their true race.
The experiment was meant to show the effects of the bias of the creators of algorithms against or in favor of certain names and how they infer race subconsciously.
​
Tools used (in my portion of the project): Excel
Project #2
Testing of an Urban Economic Theory Through a Natural Experiment
In urban economic theory, cultural aspects of certain societies can affect how people cluster. Asian countries, as compared to their Western counterparts, tend to cluster and create density at a higher rate. I wanted to see if the cultural characteristics of Asian immigrants would change the spatial makeup of the cities they immigrated to.
I compared Atlanta cities with an increasing Asian population to those that had not seen an Asian population to see if there was a causal effect.
The results showed that even though there was an increase in density, it was not likely to be a causal result.
A possible explanation for why we don't see significant density increases is that there just needs to be a longer period of time before density increases from when immigration happens.
Another possibility is that instead of new immigration from the Asian population, it is just a dissemination of the Asian population from certain areas to others as time goes on.
​
Tools Used: R, Excel
​
Please Click the Image for Full Article:
Analysis is in Section 5 of the Review
Project #3
Please Click the Image for Full Article
The Effect of Empty Stadiums During COVID-19 on Home Field Advantage in The Premier
This paper examines the effect of empty stadiums on home field advantage in the Premier League. Home field advantages are a long-held staple of sports economics that shows that having a home field advantage is in part bolstered by having a crowd cheering you on.
We can look solely at this variable during the 2019-2020 season as it is completely taken out of the equation halfway through. We will be comparing results and goals scored by home versus away teams at the beginning of the season (pre-empty stadiums) against those at the end of the season, during empty stadiums. Travel fatigue will not have changed, so it will be controlled for.
When looking at home field advantage, crowd support can play a large role in helping the home team win. Our data did not produce significant results supporting this for the season split in two by COVID. However, this could be due to outside factors that we did not consider.
​
Tools Used: Excel
Project #4
Teach Grant
Effect on High School
Success in Georgia
The model that is used to estimate the impact of Teach Grant on high schoolers' success is the Difference in Difference model.
The post variable is the mark for when the treatment begins in 2014-2017. The treatment group variable is the 25% poorest counties, which I compiled from a list of Title One Counties in 2017. "Title One" is a distinction given to a school or county when a large percentage of the system’s students are low-income. The treatment variable was a dummy variable, and I selected the 25% lowest income counties to make it into a dummy. The control would be the highest-income 25% of counties.
The results are that it is inconclusive whether you can determine if the Teach Grant has increased success rates in high schoolers. The broader implications of the results are very interesting because the individual results, like the treatment and the post variables, were great at predicting the success rate of students, so I believe that the model was so good at making the prediction, but there were some omitted variables that may be causing the data not to show the significance of the interaction term.
​
Tools Used: Stata, Excel
Please Click the Image for Full Article