Simon Jackson

This outcome was not anticipated. The original intention for correlate was to wrap `cor()`, providing the same correlation functionality but then convert the results to a data frame. I had...

Piping is a convenience of twidlr, but here are some of the more important reasons: - It provides a consistent data-frame format for all models (where many require different formats...

In the r-devel thread linked above, Yihui Xie posts a workaround: https://stat.ethz.ch/pipermail/r-devel/2013-August/067210.html Although it is not an ideal solution (see answers in thread), it may be appropriate for twidlr, which...

This would likely involve a generic function that checks predict output and coerces results to expected output. Eg checks for things like matrices, or data.frames with single values, etc.

In the case of a single set of predictions, must consider whether a vector is preferred to a data.frame of one column. By having vector or data.frame, this could lead...

After posting [question on Twitter](https://twitter.com/drsimonj/status/871858297292369921), a few notes came up about cases where this may not apply: - Spatial models. See [this](https://twitter.com/davidjayharris/status/871865780723261440) and [this](https://twitter.com/davidjayharris/status/871865975359983616). - [Predicted counts in RxCx](https://twitter.com/conjugateprior/status/871864557161074688) -...