Portfolio

Teaching Stats:

A Basic Introduction to Stats Concepts

A Bookdown that covers some introductory data science. Concepts from teaching data science to undergrads at Harvard

GIS Stats Methods:

Estimating Displacement Caused by War in Tigray

Developing methods to use satellite imagery to model the displacement of people in Tigray during their civil war caused by conflict.

Using Satellites to Estimate Population

Creating models to use NASA's VIIRS satellite data to estimate populations of 10km x 10km grid cells in countries

Using Survival Models to Estimate Coup Probabilities

Using Survival Models to estimate the probability of coups in countries.

Machine Learning Examples:

Stock Price Prediction with Transformers

An experiment using Transformer models to predict stock prices

Transformer Model Portfolio

This tutorial guides students through the use of transformer-based models for Natural Language Processing tasks such as detecting hate speech, diagnosing autism, converting text to images and vice versa, and speech processing. It also includes visualizing attention masks to understand model reasoning and explores Mamba, a potential successor to transformers.

Recent Work Experience

Research Fellow for the Director of the Weatherhead Center at Harvard

  • Research and Analysis: Conducted in-depth research on international relations and political science topics. Analyzed complex data sets and synthesized findings to support the Director's academic and policy initiatives.
  • Simplification of Complex Concepts: Translated intricate research topics into simple, comprehensible terms for broader audiences, ensuring accessibility and understanding across diverse stakeholders.
  • Publication and Presentation Support: Assisted in preparing academic papers, reports, and presentations, contributing to the Center’s research output and public engagement efforts.
  • Collaborative Projects: Worked collaboratively with other fellows and researchers, fostering a multidisciplinary approach to problem-solving and policy analysis.
  • Communication Skills: Developed strong written and verbal communication skills, essential for presenting research findings and engaging with academic and policy communities.

Teaching Fellow and Instructional Fellow, Harvard University (Gov 50 Data)

  • Course Design and Instruction: Designed and taught the core undergraduate data science course, Gov 50 Data. Created open-ended assignments guiding students through the full data analysis workflow including formulating hypotheses, data sourcing, analysis, and communicating findings.
  • Pedagogical Approach: Simplified complex concepts like linear regression by relating them to familiar math concepts and used practical, hands-on methods like scatterplot data analysis to foster intuitive understanding.
  • Student Outcomes: Empowered students, even those with limited mathematical backgrounds, to confidently apply data science skills. Many students created Shiny web applications for their resumes.
  • Recognition: Awarded multiple Certificates of Distinction in Teaching by Harvard University for excellence in teaching.

GOV 1759: Behavioural Insights and Public Policy: Nudging for Good

  • Course Support and Instruction: Assisted in teaching the undergraduate course that covers the application of behavioral economics to public policy. Helped students understand and apply complex behavioral insights to real-world policy problems.
  • Behavioral Economics Focus: Introduced students to key concepts of behavioral economics, including cognitive biases, heuristics, present bias, and social norms, and their implications for public policy.
  • Practical Application: Guided students through the design and testing of behavioral interventions to address policy issues in areas such as healthcare, finance, education, and environmental conservation. Encouraged empirical analysis and evidence-based decision-making.
  • Research Translation: Broke down complex research topics into simple, comprehensible terms for students, ensuring accessibility and effective learning. Highlighted the use of behavioral insights teams in governments worldwide to develop innovative policy solutions.
  • Consulting and Policy Relevance: Supported students in developing and presenting projects that apply behavioral economics to improve public policy, enhancing their ability to use these insights in professional contexts. Emphasized real-world applications and impact.

Featured Publications

When BLUE is not best: non-normal errors and the linear model

While it is well known that least-squares estimates are sensitive to single, unusual data points, this knowledge has not led to careful practices when using least-squares estimators. Using statistical theory and Monte Carlo simulations, we highlight the importance of using more robust estimators along with variable transformations

DOI

Preprint

Albedo as a Measure of Climate Change for Civil War Research

Although numerous studies have examined the link between climate change and civil conflict, many yield null results, possibly due to misaligned climate measurements and their real impact on agrarian communities. This study suggests using Albedo as a proxy for land's carrying capacity changes and finds a positive relationship between Albedo changes and conflict onset, highlighting the importance of nuanced environmental metrics for understanding this complex relationship.

Climate Change, Desertification, and Migration (TRUNCATED VERSION)

This study examines the relationship between desertification and human migration in Syria, using albedo as a proxy for desertification and night lights for population movement. A multi-model regression analysis reveals a significant correlation between drought and urban migration, highlighting the importance of understanding environmental migration dynamics and offering insights for policymakers on climate change adaptation and sustainable land use.

Bringing the Civil Back into Civil Military Relations

The civil-military relations literature often emphasizes the military's role in governance, but this study argues that civilian state and political parties significantly shape military strategy. Using Cox Proportional Hazards models, it shows that state and party strength better predict coup likelihood than traditional military strength indicators, highlighting the need to integrate insights from institutional studies into civil-military relations theory.

CV

You can view my CV here.