R Package to Visualize Twitter Analytics
I recently developed an R package using some of my functions to visualize descriptive stats on Twitter data. Here, ladies and gentlemen, I present “Rtwitter”.
Rtwitter Package Tutorial
I developed an R package to let you analyze and visualize Twitter data. This is a tutorial on how to use functions from the package to create some cool plots and tables.
Code - Get Instant Notification from New Jersey’s MVC with Python and Cron
Can’t you find an appointment from New Jersey’s motor vehicle department? This script lets you send an email to yourself as soon as someone cancels her appointment.
Code: Connect your Twitter Bot to Python with Twurl
So you want to code a good bot for Twitter? Here is how.
Code - Tweet Extraction from Command Line with Twarc2
I just found the easiest (and the maximalist!) way of getting tweet data using Twitter API v2. And the monthly cap is 10 million! This blog post gives you an idea what it is, how it works, and how to use twarc2.
Code - Supervised and Unsupervised Machine Learning to Estimate Polarization
This is the repo for replication materials for the tweets on Afghanistan research. It contains shell, Python and R scripts to be run in American University’s high-performance servers. You can find code to extract tweets, estimate user ideologies of Twitter users, their bot probabilities, sentiment analysis, and topic modeling (STM) here.
Research - Supervised and Unsupervised Machine Learning to Estimate Polarization
Here, I estimate user ideologies on Twitter, apply bot detection, sentiment analysis, and topic modeling on tweets from the U.S. posted during the Afghanistan withdrawal to investigate how foreign-policy polarization shapes online discussions.
Research - Estimating Conflict Effect on Growth by ArcGIS and Time-Series Analysis
I apply time-series modeling after using GIS methods and image data from NASA satellites to investigate the local effects of civil conflicts on economic growth. I convert image data to geo-coded night-time light-emissions data first, and then estimate the change of economic growth in the grid level.
Bayesian Analysis of Military Coups through MCMC Simulations
Using multilevel Bayesian logistic modeling via Markov chain Monte Carlo simulations (MCMC), this research note investigates three hypotheses on the potential determinants of military coup success; namely plotter ranking, civilian disobedience, and incumbents’ counterbalancing forces.