Welcome to Comp-Engine Time Series, a comparison engine for time-series data and time-series analysis methods.
This website opens up the results of years of work collecting and synthesizing tens of thousands of time series, and thousands of existing and newly-developed methods for measuring structure in time series. Click the banners below for an overview of why this resource might be useful for science, or to begin exploring our libraries of time-series data and Matlab-based time-series analysis code.
hctsa Matlab software for running highly comparative time-series analysis is available on github.
The key scientific article describing how this resource was used to perform highly comparative time-series analysis can be found here: Highly comparative time-series analysis: The empirical structure of time series and their methods, J. Roy. Soc. Interface, 10 (83) 20130048. Please cite this work if you use this web resource.
All (unrestricted) data used in the above article is included in this web resource, as are the hundreds of pieces of code developed for performing time-series analysis. Feel free to explore and play around with this comprehensive database of scientific data and analysis code: visualize the data, probe their interconnections, or download whole portions of the resource to analyze in more detail.
We hope to continue to develop this resource with input from the scientific community, and that it becomes a truly dynamic tool for guiding progress in time-series analysis as it occurs across the scientific disciplines. In particular we intend to incorporate functionality allowing users to upload their own data and analysis code to compare against the existing database.
In parallel with this resource, we are also developing a repository of Matlab code providing the full architecture required to compute and analyze the results of thousands of algorithms operating on specific time-series datasets. Please sign up to the mailing list (on the right), or email Ben Fulcher to express interest in a preliminary distribution of this code.
Any feedback about the website and broader initiative is much appreciated! Please provide it on the Contact page.