• Used SQL/Pyspark to manipulate large volumes of consumer/voter data and create samples of likely voters, which were then used for polling.

• Leveraged Tableau to create dashboards of the demographic and attitude breakdowns of the polling sample

• Constructed ETL pipelines for client retail stock-holder databases, later used in client-specific survey targeting

• Utilizing Python and R for web scraping thousands of newspaper editorials and radio transcripts for text analysis

• Created machine-learning models using TensorFlow and Keras that to maximize efficiency in data classification

• Applied Latent Dirichlet Allocation to create topic models of radio transcripts, creating a lexicon for further analysis