Jamie Morton
It can be easily to confuse the numerator / denominator colorings of the balances across different plotting functions. Need to make sure that this is easy.
If the tree doesn't make sense, it would be nice to throw an error for it.
Could help supplement the p-values with each of the balances. In addition, this could add additional information about the effect size, since this is a measure of classification error. https://en.wikipedia.org/wiki/Receiver_operating_characteristic
[Vega](https://vega.github.io/vega/) plots are a bit more versatile compared to bokeh plots -- it is easy to setup interactive plots that can modify the underlying data rendered. This will make it...
It would be nice to have a formal framework for multivariate response linear mixed effects https://www.jstatsoft.org/article/view/v033i02/v33i02.pdf http://sci2s.ugr.es/keel/pdf/specific/articulo/schafer_yucel02.pdf http://www.bristol.ac.uk/media-library/sites/cmm/migrated/documents/mixed-multivariate-responses.pdf http://lme4.r-forge.r-project.org/lMMwR/lrgprt.pdf https://socserv.socsci.mcmaster.ca/jfox/Books/Companion/appendix/Appendix-Multivariate-Linear-Models.pdf This will not only clean up the linear mixed effects...
It would be super cool if we could enable multiblock regression: http://onlinelibrary.wiley.com/doi/10.1002/cem.1180030104/abstract What this will ultimately allow us to do is combine multiple omics datasets together. You can imagine that...
There is a bit of recursion going on in _dendrogram.py. Would nice to completely remove it.
- [ ] Refactor return parameter suggested by @antgonza - [ ] Incorporate speed ups suggested by @wasade
The helper methods for calculating `depth` needs to be fixed.
Namely, we need to introduce - [ ] `Balance` type - [ ] `Hierarchy` type