DATA

“Audible Reckoning: How top political podcasters spread unsubstantiated and false claims”

Provided research assistance for the Brookings Institution during my Fall 2022 internship. Extensive research summarized in The New York Times.

The rightmost plots are the resulting graphs published in the extensive report. The relation of topic frequency fact checked by party was not included in the final report, as many of the top political podcasters studied are ideologically conservative by nature. Adding a “conservative” variable (as illustrated in the top left plot) would saturate the relatively small “liberal” variable found in the dataset. Special formatting and content changes to the color and text were noted in the resulting plots by the senior researchers of this report.

I created a grouped/ungrouped bar graph using open-source visualization packages '“ggplot” and “ggbrookings” in R to analyze the frequency of podcasting topics fact checked by party (top left) and frequency of people fact checked by podcast hosts (bottom left). For each plot, I added a two-facet feature to distinguish the chart when “host refutes false claims” and “host shares false claims." These preliminary plots (leftmost) were used to determine what trends were available in our dataset before identifying which variables were necessary to include.

“Will Brazil’s elections be a victory for democracy?”

Replicated Google Search Trends results in a Brookings Institution report.