PPOL564 | Data Science 1 | Foundations Week 7 Data Visualization

# PPOL564 | Data Science 1 | Foundations Week 7 Data Visualization 
###  Prof. Eric Dunford  ◆  Georgetown University  ◆  McCourt School of Public Policy  ◆  <a href="mailto:eric.dunford@georgetown.edu" class="email">eric.dunford@georgetown.edu</a>

---

<div class="slide-footer"> 
PPOL564 | Data Science 1 | Foundations

&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;

Week 7

&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;&emsp;

Data Visualization

</div>

---

# The Components of Data Visualization

---

### What do you see?

---

### Mapping data to space

---

### Aesthetics

.center[<img src="Figures/common-aesthetics-1.png", width=500>]
 
.center[<img src="Figures/basic-scales-example-1.png", width=500>]

---

### Color as a tool to distinguish

---

### Color as a tool to represent values

---

### Color as a tool to highlight

---

### Presentation as information

---

### Presentation as distortion

---

### The data type drives the visualization decisions

Think carefully about what you're trying to convey and what information you're using to make your point.

| Data Type   | Example |  Scale  |
|-------------|---------|---------|
| Numerical   | `1.3`, `800`, `10e3` | Continuous | 
| Integer     | `1`, `2`, `3` | Discrete | 
| Categorical | `"dog"`, `"Nigeria"`, `"A"` | Discrete |
| Ordered     | `"Small"`, `"Medium"`, `"Large"` | Discrete |
| Dates/Time  | `2009-01-02`, `5:32:33` | Continuous |

]

---

### The data type drives the visualization decisions

.center[
.center[<img src="Figures/amounts-1.png", width=500>]
.center[<img src="Figures/proportions-1.png", width=500>]
.center[<img src="Figures/single-distributions-1.png", width=500>]
.center[<img src="Figures/basic-scatter-1.png", width=500>]
]

---

class:newsection

# Grammar of Graphics

---

.pull-left[.center[<img src = "Figures/ggplot-hex.png"> &#8776; <img src = "Figures/plotnine_logo.png">]]

.pull-right[
`plotnine` is an emulator for the powerfu `ggplot2` graphics package in `R`. `ggplot2` offers a flexible and intuitive graphics language capable of building sophisticated graphics.

`plotnine`/`ggplot2` has a **special syntax** that we'll have to get used to, _but_ once we understand the basics, we'll be able to produce some advanced and sophisticated graphics with ease!
]

---

.pull-left[.center[<img src = "Figures/ggplot-hex.png"> &#8776; <img src = "Figures/plotnine_logo.png">]]

`plotnine`/`ggplot2` is based on a **grammar of graphics**. In essence, you can build every graph from the same components that follow the same intuitive naming conventions.

Every graph is composed of

1. a **dataset**

2. **coordinate system**

2. **mappings** &rarr; the variables we're aiming to visualize

3. **geom**etric expressions of how the data should be projected onto a space

]

---

`+`

`<GEOM_FUNCTION>`(mapping = aes(`<MAPPINGS>`))
]

`+`

`<GEOM_FUNCTION>`(mapping = aes(`<MAPPINGS>`))

`+`

`<GEOM_FUNCTION>`(mapping = aes(`<MAPPINGS>`))

`+`

`<GEOM_FUNCTION>`(mapping = aes(`<MAPPINGS>`))

`$$\vdots$$`

]

---

`+`

`<GEOM_FUNCTION>`(mapping = aes(`<MAPPINGS>`))
]

`+`

`<SCALE_FUNCTION>`(mapping = aes(`<MAPPINGS>`))

`+`

`<THEME_FUNCTION>`(mapping = aes(`<MAPPINGS>`))

`+`

`<FACET_FUNCTION>`(mapping = aes(`<MAPPINGS>`))

`$$\vdots$$`

]

---

### One variable?

.center[
| Expression | Function | |
|----|----|-----|
| Area | `geom_area()` | <img src = "Figures/geom_area.png"> | 
| Density | `geom_density()` | <img src = "Figures/geom_density.png"> | 
| Dots | `geom_dotplot()` | <img src = "Figures/geom_dotplot.png"> | 
| Frequencies | `geom_freqpoly()` | <img src = "Figures/geom_freqplot.png"> | 
| Histogram | `geom_histogram()` | <img src = "Figures/geom_histogram.png"> | 
]

---

### Two variables?

.center[
| Expression | Function | |
|----|----|-----|
| Continuous Points | `geom_point()` | <img src = "Figures/geom_point.png"> | 
| Continous Lines | `geom_line()` | <img src = "Figures/geom_line.png"> | 
| Discrete Counts | `geom_count()` | <img src = "Figures/geom_count.png"> | 
| Continuous and Discrete Distributions | `geom_boxplot()` | <img src = "Figures/geom_boxplot.png"> | 
| Densities | `geom_hex()` | <img src = "Figures/geom_hex.png"> | 
]

---

### Three variables?

.center[
| Expression | Function | |
|----|----|-----|
| Densities | `geom_contour()` | <img src = "Figures/geom_contour.png"> | 
| Intensities | `geom_tile()` | <img src = "Figures/geom_tile.png"> | 
| Intensities | `geom_raster()` | <img src = "Figures/geom_raster.png"> | 
| Spatial | `geom_map()` | <img src = "Figures/geom_map.png"> | 
]

See [`plotnine`'s documentation website](https://plotnine.readthedocs.io/en/stable/index.html) for additional guidance and tips on using the API.

---

### Function Types in `ggplot2`/`plotnine`

| Type | Function Header | Description |
|------|-----------------|-------------|
|  Generate layers from data   |  `geom_` | Use a geom function to represent data points, use the geom’s aesthetic properties to represent variables. Each function returns a layer. |
|   Construct statistics layers   |  `stat_` | A stat builds new variables to plot (e.g., count, prop) |
|   Change mapping characteristics   |  `scale_` | Scales map data values to the visual values of an aesthetic. To change a mapping, add a new scale. |
|   Generate subplots   |  `facet_` | Facets divide a plot into subplots based on the values of one or more discrete variables.  |
|   Alter the plots theme   |  `theme_` | Change the aesthetics of the plot background and feature (e.g. axes, text, grid lines, etc.)  |