CRT Data Visualisation Workshop

---

### What to expect

.pull-right[
#### Day 2
- ggplot2 extensions
- Other tools and things you should know about 
- Data viz challenge
- Presentations
]

---

# Introductions
---
### Me
- Cohort 1 CRT student
- Interested in data visualisation
- Currently work for a drug discovery startup based in Toronto

![](imgs/me1.png)
---
### Me 
- Cohort 1 CRT student
- Interested in data visualisation
- Currently work for a drug discovery startup based in Toronto

![](imgs/me2.png)
---
### Me 
- Cohort 1 CRT student
- Interested in data visualisation
- Currently work for a drug discovery startup based in Toronto

![](imgs/me3.png)
---
## Why you should care about data visualisation

Find hidden patterns in data

Charts change minds

Charts change minds

Good data viz can save lives

Good data viz can save lives

Good data viz can save lives

---
## Why you should care about data visualisation

Set yourself apart from others

---
## What makes a good plot

It depends who you ask...

.pull-right[
- Influential voice on data visualisation
- Coined the terms “chart junk” and “data-ink ratio”
- “The number of information carrying (variable) dimensions depicted should not exceed the number of dimensions in the data”
]

---
## What makes a good plot

It depends who you ask...

.pull-left[
.center[
Florence Nightingale 
![flo1](imgs/florence.jpg) 
Pioneer of colourful, clear visualisations aimed at a popular audience
]
]

# Tips & Tricks
---
## Choosing the correct chart type
Considerations
- What type of data are you working with - distribution or trend in a single variable or a relationship or comparison between multiple variables 
- What is the audience for the visualisation - lab meeting, conference presentation, thesis/publication
- What question do you want the plot to answer
--
.center[
 
*Choose a chart type that will visualise the message you want to convey in the most effective and obvious way*
]

---
## Colour ♥️💙💚💛💜

Palette types

.center[
 ![](imgs/palette_types.png)
]
---
## Colour ♥️💙💚💛💜
Colour blindness
.center[
![col_blind](imgs/col_blindness.png)
]
---
## Colour ♥️💙💚💛💜
Colour blindness
.pull-left[
Quantitative (sequential)
![](imgs/viridis1.png)
[viridis palettes](https://cran.r-project.org/web/packages/viridis/vignettes/intro-to-viridis.html)
]

.pull-right[
Quantitative (diverging)
![](imgs/seurat1.png)
[Seurat purple yellow](http://satijalab.org/seurat/reference/custompalette)
 
Qualitative
![](imgs/pal_martin1.png)
[Martin Kryzwinski palettes](https://mk.bcgsc.ca/colorblind/palettes.mhtml)
]
---
## Colour ♥️💙💚💛💜
Colour blindness
.pull-left[
Quantitative (sequential)
![](imgs/viridis2.png)
[viridis palettes](https://cran.r-project.org/web/packages/viridis/vignettes/intro-to-viridis.html)
]

.pull-right[
Quantitative (diverging)
![](imgs/seurat2.png)
[Seurat purple yellow](http://satijalab.org/seurat/reference/custompalette)
 
Qualitative
![](imgs/pal_martin2.png)
[Martin Kryzwinski palettes](https://mk.bcgsc.ca/colorblind/palettes.mhtml)
]

---
## Colour ♥️💙💚💛💜
Colour blindness
.pull-left[
Quantitative (sequential)
![](imgs/viridis3.png)
[viridis palettes](https://cran.r-project.org/web/packages/viridis/vignettes/intro-to-viridis.html)
]

.pull-right[
Quantitative (diverging)
![](imgs/seurat3.png)
[Seurat purple yellow](http://satijalab.org/seurat/reference/custompalette)
 
Qualitative
![](imgs/pal_martin3.png)
[Martin Kryzwinski palettes](https://mk.bcgsc.ca/colorblind/palettes.mhtml)
]
---
## Colour ♥️💙💚💛💜
Be consistent 
![hel1](imgs/helen1.png)![hel2](imgs/helen2.png) 
[*Horkan et al, 2025*](https://www.nature.com/articles/s41467-025-57168-z)

---
## Colour ♥️💙💚💛💜
Be consistent... but not like this...
![](imgs/bad_colour.png)

---
## Ordering variables
.center[
![reord](imgs/reorder_variables_1.png)
]

---
## Ordering variables
.center[
![reord](imgs/reorder_variables_2.png)
]

---
## Avoid rotated labels

---
## Use annotations and labels

![](imgs/annot.png)
[*Stokes et al, 2022*](https://arxiv.org/abs/2208.01780)

--
![](imgs/annot1.png)

---
## Themes

``` r
library(ggplot2)
library(palmerpenguins)
data(package = 'palmerpenguins')

ggplot(penguins, 
       aes(x = bill_length_mm, 
           y = bill_depth_mm, 
           col = sex)) +
  geom_point() +
  facet_wrap(~species)
```
]

---
## Themes

``` r
library(ggplot2)
library(palmerpenguins)
data(package = 'palmerpenguins')

ggplot(penguins, 
       aes(x = bill_length_mm, 
           y = bill_depth_mm, 
           col = sex)) +
  geom_point() +
  facet_wrap(~species) +
* theme_minimal(base_size = 14)
```
]

---
## Themes

``` r
library(ggplot2)
library(palmerpenguins)
data(package = 'palmerpenguins')

ggplot(penguins, 
       aes(x = bill_length_mm, 
           y = bill_depth_mm, 
           col = sex)) +
  geom_point() +
  facet_wrap(~species) +
* theme_bough()
```
]

---
## Interactivity and animation

.pull-left[
- Plots don’t have to be static
- Consider animations for presentations
- Link to interactive apps in your publications/thesis
]

---

# ggplot2
---
## ggplot2

![](imgs/ggplot.png)
---
## The grammar of graphics
.pull-left[
- Language proposed by Leland Wilkinson in 2005
- Foundation for building almost any type of plot
- Plots are built up in sequential layers
- Implemented in ggplot2 package as well as other packages
]

``` r
ggplot(data = mpg, 
       aes(x = hwy, y = displ, colour = class)) 
```
]

.right-plot[
<img src="day1_files/figure-html/gg1-out-1.png" width="100%" />
]
---
## The anatomy of a ggplot

``` r
ggplot(data = mpg, 
       aes(x = hwy, y = displ, colour = class)) +
  geom_point()
```
]

.right-plot[
<img src="day1_files/figure-html/gg2-out-1.png" width="100%" />
]
---
## The anatomy of a ggplot

``` r
ggplot(data = mpg, 
       aes(x = hwy, y = displ, colour = class)) +
  geom_point() +
  scale_colour_manual(values = rainbow(7))
```
]

.right-plot[
<img src="day1_files/figure-html/gg3-out-1.png" width="100%" />
]
---
## The anatomy of a ggplot

``` r
ggplot(data = mpg, 
       aes(x = hwy, y = displ, colour = class)) +
  geom_point() +
  scale_colour_manual(values = rainbow(7)) +
  labs(x = 'Highway miles per gallon',
       y = 'Engine displacement (litres)')
```
]

---
## The anatomy of a ggplot

---
## A more relevant example
How to get from this...
![](imgs/marker_ex_default.png)

---
## A more relevant example
To this...
![](imgs/marker_enhanced.png)
---
## Basic plot

``` r
marker_df <- read.csv('https://raw.githubusercontent.com/Sarah145/CRT-data-viz/refs/heads/main/data/bm_marker_genes.csv')

ggplot(marker_df, 
       aes(x = gene, y = celltype, colour = scaled_avg_exp, size = per_exp)) +
  geom_point() +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5))
```

---
## Enhanced plot
Setup...

``` r
library(ggtext)

gene_order <- c('SDC1','MME','CD19','CD79A','MS4A1','KIT','CD34','CLC','GYPA','PPBP','ITGA2B','CD33','MPO','CTSG','AZU1','LYZ','S100A8','S100A9','CD14','ITGAM','FCGR3A','FCER1G','FCER1A','CLEC10A','IL3RA','IRF8','CD4','IL7R','CD3E','CCR7','CD8A','CCL5','GNLY','KLRB1','NKG7','CXCL12','VCAM1')

celltype_cols <- c(Plasma="#333333", CLP="#B2EF9B", Pre.B="#3A7219", B="#1B360C", HSC.MPP="#F7D760", MEP="#B84F09", Late.Eryth="#6D2F05", Megakaryocytes="#F7A660", GMP="#DE9197", CD14.Mono="#AD343E", CD16.Mono="#BD0000", cDC="#FF0378", pDC="#FC5E03", CD4.N="#015073",CD4.M="#96CBFE", CD8.N="#16B8B2", CD8.M="#946CD5", NK="#4D1282", Mesenchymal="#702127")

gradient_cols <- colorRampPalette(c('#190C3E', '#87216B', '#E55C30', '#F7D340'))

y_labs <- paste0("", names(celltype_cols), "")

head(y_labs)
```

```
## [1] "Plasma" 
## [2] "CLP" 
## [3] "Pre.B" 
## [4] "B" 
## [5] "HSC.MPP"
## [6] "MEP"
```

---
## Enhanced plot

``` r
ggplot(marker_df, 
       aes(x = gene, y = celltype, colour = scaled_avg_exp, size = per_exp)) +
  geom_point() + 
  scale_x_discrete(limits = gene_order) +
  scale_y_discrete(limits = rev(names(celltype_cols)), labels = rev(y_labs)) +
  scale_colour_gradientn(colours = gradient_cols(100)) +
  scale_size(labels = scales::percent, range = c(0.5, 4)) +
  labs(x = NULL, y = NULL, colour = 'Avg. expression\n(scaled 0-1)', size = '% cells\nexpressing\ngene') +
  guides(colour = guide_colourbar(order = 1, barlength = unit(6, 'lines')), size = guide_legend(order = 2)) +
  theme_linedraw(base_size = 14) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5),
        axis.text.y = element_markdown(face = 'bold'),
        legend.title = element_text(size = 12, vjust = 0.5),
        legend.box.margin = margin(t = 45))
```

---

# BYOD