Hourly Subway Station Flows

This article discusses the visualization of hourly passenger flows in the New York City Subway system using 424 animated pie charts, despite the known drawbacks of pie charts for comparing proportions. The author uses MTA origin-destination estimates aggregated over all Mondays in 2025 to calculate in-and-out flows at 424 stations. The pie charts show departures in yellow and arrivals in purple, with circle size representing total volume. The article also covers data processing techniques using parquet files, DuckDB, and duckplyr to handle a 121-million-row dataset, reducing a 17GB CSV to a 1.5GB parquet file. It touches on transit data analysis, data visualization best practices, and technical data manipulation.