36-613: Data Visualization

class: center, middle, inverse, title-slide

.title[
# 36-613: Data Visualization
]
.subtitle[
## Introduction and the Grammar of Graphics
]
.author[
### Professor Ron Yurko
]
.date[
### 8/29/2022
]

---

## Who am I?

.pull-left[

*   Assistant Teaching Professor

*   Finished Phd in Statistics @ CMU in May 2022

*   Previously BS in Statistics @ CMU in 2015

*   Research interests: statistical genetics, selective inference, clustering, sparsity, statistics in sports

*   Industry experience: briefly worked in finance before returning to grad school and also as data scientist in professional sports
]

.pull-right[

![Osborn](https://www.meme-arsenal.com/memes/985cf50f0f3a6ad462d072513394e0a0.jpg)

]

---
name: quartet

## Do these [datasets](http://www.thefunctionalart.com/2016/08/download-datasaurus-never-trust-summary.html) have anything in common?

.center[![](https://d2f99xq7vri1nk.cloudfront.net/DataDino-600x455.gif)]

---

## __Always visualize your data__ before analyzing it!

.center[![](https://www.researchgate.net/profile/Arch_Woodside2/publication/285672900/figure/fig4/AS:305089983074309@1449750528742/Anscombes-quartet-of-different-XY-plots-of-four-data-sets-having-identical-averages.png)]

Need to understand the interplay between graphics and statistical inference

---

## Course Structure (READ THE SYLLABUS):

#### Monday / Wednesday = Lecture

+ Will include example code, all slides and additional `R` demos posted on https://cmu-36613.netlify.app/

+ Want class discussion - so __participate and ask questions__

#### 5 Weekly Homework Assignments due Wednesdays by 11:59 PM EST

+ Posted Wednesday mornings and due one week later

#### Two Graphics Critique / Replication of Data Viz in the Wild

+ Submit two graphics critique / replications (with pseudocode) of data visualizations
you find in the wild (see syllabus for due dates)

+ Both data visualizations must be from a __recent source that was posted online within the past month__

#### Group EDA Project due Friday October 14th by 11:59 PM EST

+ Each group will write an IMRD report and present their work in 36-611

---

## Course Objectives (READ THE SYLLABUS):

### Learn useful principles for making appropriate statistical graphics.

### Critique existing graphs and remake better ones.

### Visualize statistical analyses to facilitate communication.

### Pinpoint the statistical claims you can/cannot make from graphics.

### Practice tidy data manipulation in `R` using the `tidyverse`

### Practice reproducible workflows with `RMarkdown`

---

## What do I mean by `tidy` data?

Data are often stored in __tabular__ (or matrix) form:

```
## # A tibble: 5 × 8
##   species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g sex     year
##   <fct>   <fct>              <dbl>         <dbl>             <int>       <int> <fct>  <int>
## 1 Adelie  Torgersen           39.1          18.7               181        3750 male    2007
## 2 Adelie  Torgersen           39.5          17.4               186        3800 female  2007
## 3 Adelie  Torgersen           40.3          18                 195        3250 female  2007
## 4 Adelie  Torgersen           NA            NA                  NA          NA <NA>    2007
## 5 Adelie  Torgersen           36.7          19.3               193        3450 female  2007
```

+ Each row `==` unit of observation, e.g., penguins

+ Each column `==` variable/measurement about each observation, e.g., `flipper_length_mm`

+ Known as a `data.frame` in base `R` and `tibble` in the `tidyverse`

+ Two main variable types: quantitative and categorical

__How do we convert data into visualizations?__

---

## [The Grammar of Graphics](https://link.springer.com/book/10.1007/0-387-28695-0) - by Leland Wilkinson

.pull-left[

All plots can be broken down into core components

1. __data__

2. __geometries__: type of geometric objects to represent data, e.g., points, lines

3. __aesthetics__: visual characteristics of geometric objects to represent data, e.g., position, size

4. __scales__: how each aesthetic is converted into values on the graph, e.g., color scales

5. __stats__: statistical transformations to  summarize data, e.g., counts, means, regression lines

6. __facets__: split data and view as multiple graphs

7. __coordinate system__: 2D space the data are projected onto, e.g., Cartesian coordinates

]

.pull-right[

[Hadley Wickham](http://hadley.nz/) [expanded upon this](http://vita.had.co.nz/papers/layered-grammar.pdf) with  [`ggplot2`](https://ggplot2.tidyverse.org/)

1. `data`

2. `geom`

3. `aes`: mappings of columns to geometric objects

4. `scale`: one scale for each `aes` variable

5. `stat`

6. `facet`

7. `coord`

8. `labs`: labels/guides for each variable and other parts of the plot, e.g., title, subtitle, caption

9. `theme`: customization of plot layout

]

---

## Start with the `data`

.pull-left[

Access `ggplot2` from the `tidyverse`:

```r
library(tidyverse)
*ggplot(data = penguins)
```

Or equivalently using `%>%`:

```r
*penguins %>%
  ggplot()
```

]

.pull-right[

__Nothing is displayed__

]

---

## Add geometric object with columns mapped to aesthetics

.pull-left[

+ Use the `+` operator in `ggplot` to add layers

+ Map `bill_length_mm` to x-axis and `bill_depth_mm` to y-axis

```r
penguins %>%
* ggplot(aes(x = bill_length_mm,
*            y = bill_depth_mm)) +
* geom_point()
```

+ NOTE we are implicitly saying:

```r
penguins %>%
* ggplot(mapping = aes(x = bill_length_mm,
                       y = bill_depth_mm)) + 
  geom_point() 
```

]

.pull-right[

__And now we have a scatterplot!__

]

---

## Modify scale, add statistical summary, and so on...

.pull-left[

+ Adjust global aesthetics outside of `aes`

```r
penguins %>%
  ggplot(aes(x = bill_length_mm,
             y = bill_depth_mm)) + 
  # Adjust alpha of points
* geom_point(alpha = 0.5) +
  # Add smooth regression line
* stat_smooth(method = "lm") +
  # Flip the x-axis scale
* scale_x_reverse() +
  # Change title & axes labels 
* labs(x = "Bill length (mm)",
*      y = "Bill depth (mm)",
*      title = "Clustering of penguins bills") +
  # Change the theme:
* theme_bw()
```

]

.pull-right[

__You will be covering more basics in HW1__

]

---

## In the beginning...

#### Michael Florent van Langren published the first (known) statistical graphic in 1644

+ Plots different estimates of the longitudinal distance between Toledo, Spain and Rome, Italy

+ i.e., visualization of collected data to aid in estimation of parameter

---

## [John Snow](https://www.theguardian.com/news/datablog/2013/mar/15/john-snow-cholera-map) Knows Something About Cholera

---

## [Charles Minard's](https://www.datavis.ca/gallery/minard/minard.pdf) Map of Napoleon's Russian Disaster

---

## [Florence Nightingale's](https://www.datavis.ca/gallery/flo.php) Rose Diagram

---

## [Milestones in Data Visualization History](https://friendly.github.io/HistDataVis/)

---

## How to Fail this Class:

---

## [Edward Tufte's](https://www.edwardtufte.com/tufte/) Principles of Data Visualization

Graphics: visually display measured quantities by combining points, lines, coordinate systems, numbers, symbols, words, shading, color

#### Goal is to show data and/or communicate a story!

+ Induce viewer to think about substance, __not graphical methodology__

+ Make large, complex datasets more coherent

+ Encourage comparison of different pieces of data

+ __Describe, explore, and identify relationships__

+ __Avoid data distortion and data decoration__

+ Use consistent graph design

#### Avoid graphs that lead to misleading conclusions!

---

#### [Think twice before you spiral](https://junkcharts.typepad.com/junk_charts/nyt/)

---
class: center, middle

# Next time: 1D categorical data