Last major update, Fall 2018: Early work on this data resource was funded by an NSF Career Award 0237918, and has been funded through NSF IIS-1161997 II and NSF IIS 1510741.

News! The Hexagon ML/UCR Time Series Anomaly Detection datasets are here.

UCR Time Series Classification Archive

We suggest you begin by reading the briefing document in PDF or PowerPoint, which also contains the password. Then you can download the entire archive (about 260 MB in zipped format).

We strongly recommend you to read this paper for a detailed discussion of how the community can best benefit from the archive. If you are interested in multivariate (multi-dimensional) time series, there is a collection of thirty such datasets archived here: www.timeseriesclassification.com, read this paper for details.

Please reference as: Hoang Anh Dau, Eamonn Keogh, Kaveh Kamgar, Chin-Chia Michael Yeh, Yan Zhu, Shaghayegh Gharghabi , Chotirat Ann Ratanamahatana, Yanping Chen, Bing Hu, Nurjahan Begum, Anthony Bagnall , Abdullah Mueen, Gustavo Batista, & Hexagon-ML (2019). The UCR Time Series Classification Archive. URL https://www.cs.ucr.edu/~eamonn/time_series_data_2018/

Fun fact: We estimate that to build this archive we needed to invoke DTW 321,053,852 times at testing time, and 61,041,100,000,000 times at training time (learning the warping window width). Yet, to this day there are still papers that bemoan that the “DTW is too slow…” If you think that, check out the UCR Suite.

title = {The UCR Time Series Classification Archive},
author = {Dau, Hoang Anh and Keogh, Eamonn and Kamgar, Kaveh and Yeh, Chin-Chia Michael and Zhu, Yan
and Gharghabi, Shaghayegh and Ratanamahatana, Chotirat Ann and Yanping and Hu, Bing
and Begum, Nurjahan and Bagnall, Anthony and Mueen, Abdullah and Batista, Gustavo, and Hexagon-ML},
year = {2018},
month = {October},
note = {\url{https://www.cs.ucr.edu/~eamonn/time_series_data_2018/}}

You can download the entire spreadsheet displayed below in CSV format or Excel format.