This is a companion discussion topic for the original entry at:
The world of time series and financial analysis in R has diverged in terms of the tooling used in day to day work. xts and quantmod currently reign supreme with a powerful arsenal of tools available to analyze time-based data. However, this data often requires manipulation and visualization using the tools of the tidyverse. Moreover, the grouped analysis capabilities of dplyr are powerful features that are not available to the xts world. To reconcile these differences, we propose a ‘time-aware tibble’ that builds off of tibble and the tidyverse as a whole, but with a knowledge of the column to be used as the index for time-based subsetting, rolling analysis, period-based summaries, and endless other applications. The tibbletime package is the first step towards this. Eventually, this package will serve as the foundational data structure for a number of other packages focused on performance analysis, portfolio construction, and forecasting. Currently, combined with the tidyquant package, the two allow for a more seamless transition between xts and the tidyverse, but the eventual goal is to perform time series analysis completely in the tidyverse.
Materials: The future of time series and financial analysis in the tidyverse
GitHub: Davis Vaughan
Davis Vaughan - Manager of Software Developmenta
Data science consultant interested in opportunities at the intersection of R, finance, and data analytics. Co-author of the R packages managed by Business Science, LLC including: