dtm2451 / dittoviz Goto Github PK

Color blindness friendly visualization of any data in a data.frame format. Like dittoSeq, but not omics-structure restricted

License: Other

R 99.83% TeX 0.17%

color-blindness ggplot2 visualization

dittoviz's Introduction

dittoViz

A comprehensive visualization toolkit built with coders of all skill levels and color-vision impaired audiences in mind. It allows creation of finely-tuned, publication-quality figures from single function calls.

Spawned out of the scRNAseq-focused dittoSeq-package, dittoViz contains most of it's visualization machinery with none of the added scRNAseq-related dependencies.

Visualizations include scatter plots, compositional bar plots, violin, box, and ridge plots, and more. Customization ranges from size and title adjustments to discrete-group circling and labeling, hidden data overlay upon cursor hovering via ggplotly() conversion, and many more, all with simple, discrete inputs.

Color blindness friendliness is powered by legend adjustments (enlarged keys), and by allowing the use of shapes or letter-overlay in addition to the carefully selected dittoColors().

Getting Started

Installation:

install.packages("dittoViz")

Then there are a few options for getting aquainted with the package:

Read and follow along with the vignette available here!
View documentation and run examples inside R with, e.g. ?yPlot or example("yPlot").

Visualization functions are:

scatterPlot()
scatterHex()
yPlot()
barPlot()
freqPlot()

Helper functions are:

dittoColors()
colLevels()

Long-Term plans for dittoViz

Carry over as much functionality from dittoSeq as makes sense for targeting data stored in a data.frame-type object (rather than in Seurat/SingleCellExperiemnt/SummarizedExperiment objects).

Much of this work is now complete! ... But there are still a few missing bits, the biggest of which are dittoDotPlot() and dittoHeatmap() counterparts.

Implement dittoViz inside of a dittoSeq-v2.0 after it gets accepted into Bioconductor.
Python counterpart: a similarly named, largely equivalent, python companion which would target pandas.DataFrames using the plotnine ggplot-mimic and as similar syntax as possible.

dittoviz's People

Contributors

Stargazers

Watchers

Forkers

j-andrews7

dittoviz's Issues

jitter of yPlot with multivar.aes="color" doesn't align properly

Current look of

library(dittoViz)
example("dittoExampleData", echo = FALSE)
yPlot(example_df, c("gene1", "gene2"), "groups", multivar.aes = "color")

It should have the jitter dots aligned with the boxplots and violin plots.

Likely just needs a group aes fix.

`extra.vars` functionality a la `dittoSeq`

This functionality is convenient in a lot of ways. Looking at dittoSeq, the data collection is pretty complicated. I haven't toyed with it here, but I assume it'd just be tweaking the .data_adjust_scatter function to pop them on as well.

I know one can use data.out to get the full data, but it being embedded in the plot object was handy in its own way.

Consider adding `do.hover` functionality for ridgeline plots

This could be hacked together directly with plotly using split violin plots.

Line chart

Would be nice to have, potentially mixed with the scatterPlot functionality.

`size.by` & `opacity.by` parameters for jitter/scatter plots

To encode yet another piece of data in the plot.

Ability to annotate points via `annotate.by` & `annotate.rows`

As we already discussed, this could be a nice complement to do.label.

`multivar.split.dir` clarification in `scatterPlot`

Is this parameter functional at all since multiple vars can't be given to x.by or y.by?

Support for point fill and border thickness in scatterPlot

E.g.

df <- data.frame(id=runif(12), x=1:12, y=runif(12))
ggplot(df, aes(x=x, y=y)) + 
  geom_point(aes(fill=id, size=id), colour="black", shape=21, stroke = 2)

This would require some adjustments to how color.by is interpreted (to switch it to populate the fill aes instead) and addition of parameters to control stroke width and color (when fill is populated rather than color by color.by). Potentially confusing changes, as we'd only want it to work with the open shapes (21-25).

Would be interested in how you'd prefer something like this be implemented.

`show.grid.lines` not functional for `scatterPlot`

E.g. scatterPlot(example_df, x.by = "PC1", y.by = "PC2", show.grid.lines = FALSE):

Ability to add multiple horizontal and vertical lines to `scatterPlot`

E.g. take a vector of numerics as input. Being able to add vertical lines as well would be nice. Could do it with a x.add.lines and y.add.lines parameters along with matching linetype and color for each axis.

Horizontal lines are already available in yPlot, and vertical lines make less sense there and would be tougher to implement.