This paper introduces the dtw package for R, which provides a comprehensive solution for computing and visualizing dynamic time warping (DTW) alignments. DTW is a technique for comparing time series by aligning them while accounting for local compression and stretching. The package unifies various algorithms and constraints, allowing users to compute alignments with flexibility in continuity constraints, restriction windows, endpoints, and local distance definitions. It also offers functions for visualizing alignments and constraints using classic diagram types. The paper outlines the DTW algorithm, explaining how it computes the optimal alignment between two time series by minimizing the accumulated distortion. It discusses different step patterns, windowing constraints, and normalization methods, which influence the alignment process. The package supports both global and partial alignments, including open-end and subsequence matching. The paper also covers multivariate time series, the use of custom distance functions, and efficient computation of multiple alignments. Examples demonstrate how to use the dtw function, visualize results, and interpret alignment outputs. The package is available on CRAN and offers extensive customization options for researchers and practitioners working with time series data.This paper introduces the dtw package for R, which provides a comprehensive solution for computing and visualizing dynamic time warping (DTW) alignments. DTW is a technique for comparing time series by aligning them while accounting for local compression and stretching. The package unifies various algorithms and constraints, allowing users to compute alignments with flexibility in continuity constraints, restriction windows, endpoints, and local distance definitions. It also offers functions for visualizing alignments and constraints using classic diagram types. The paper outlines the DTW algorithm, explaining how it computes the optimal alignment between two time series by minimizing the accumulated distortion. It discusses different step patterns, windowing constraints, and normalization methods, which influence the alignment process. The package supports both global and partial alignments, including open-end and subsequence matching. The paper also covers multivariate time series, the use of custom distance functions, and efficient computation of multiple alignments. Examples demonstrate how to use the dtw function, visualize results, and interpret alignment outputs. The package is available on CRAN and offers extensive customization options for researchers and practitioners working with time series data.