This blog post explores the analysis of subway ridership data in New York using time series decomposition techniques, specifically the MSTL method. It highlights the challenges in interpreting raw time series data, advocates for using smoothing techniques to clarify trends, and explains how to extract valuable insights through component analysis of ridership trends, seasonality, and residuals. The post also compares different subway lines based on long-term ridership changes and discusses the strengths and limitations of the MSTL approach.