Lec 11 - seaborn

class: center, middle, inverse, title-slide

# Lec 11 - seaborn
## <br/> Statistical Computing and Computation
### Sta 663 | Spring 2022
### <br/> Dr. Colin Rundel

---

exclude: true

```python
import numpy as np
import matplotlib as mpl
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns

plt.rcParams['figure.dpi'] = 200

penguins = sns.load_dataset("penguins")
```

```r
knitr::opts_chunk$set(
  fig.align="center",
  cache=FALSE
)
```

```r
local({
  hook_err_old <- knitr::knit_hooks$get("error")  # save the old hook
  knitr::knit_hooks$set(error = function(x, options) {
    # now do whatever you want to do with x, and pass
    # the new x to the old hook
    x = sub("## \n## Detailed traceback:\n.*$", "", x)
    x = sub("Error in py_call_impl\$.*?\$\\: ", "", x)
    hook_err_old(x, options)
  })
  
  hook_warn_old <- knitr::knit_hooks$get("warning")  # save the old hook
  knitr::knit_hooks$set(warning = function(x, options) {
    x = sub("<string>:1: ", "", x)
    hook_warn_old(x, options)
  })
})
```

---

## seaborn

> Seaborn is a library for making statistical graphics in Python. It builds on top of **matplotlib** and integrates closely with **pandas** data structures.
>
> Seaborn helps you explore and understand your data. Its plotting functions operate on dataframes and arrays containing whole datasets and internally perform the necessary semantic mapping and statistical aggregation to produce informative plots. Its dataset-oriented, declarative API lets you focus on what the different elements of your plots mean, rather than on the details of how to draw them.

.small[

```python
import matplotlib.pyplot as plt
import seaborn as sns
```
]

.small[

```python
penguins = sns.load_dataset("penguins")
penguins
```

```
##     species     island  bill_length_mm  ...  flipper_length_mm  body_mass_g     sex
## 0    Adelie  Torgersen            39.1  ...              181.0       3750.0    Male
## 1    Adelie  Torgersen            39.5  ...              186.0       3800.0  Female
## 2    Adelie  Torgersen            40.3  ...              195.0       3250.0  Female
## 3    Adelie  Torgersen             NaN  ...                NaN          NaN     NaN
## 4    Adelie  Torgersen            36.7  ...              193.0       3450.0  Female
## ..      ...        ...             ...  ...                ...          ...     ...
## 339  Gentoo     Biscoe             NaN  ...                NaN          NaN     NaN
## 340  Gentoo     Biscoe            46.8  ...              215.0       4850.0  Female
## 341  Gentoo     Biscoe            50.4  ...              222.0       5750.0    Male
## 342  Gentoo     Biscoe            45.2  ...              212.0       5200.0  Female
## 343  Gentoo     Biscoe            49.9  ...              213.0       5400.0    Male
## 
## [344 rows x 7 columns]
```
]

---

## Basic plots

.pull-left[

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", 
  y="bill_depth_mm"
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-3-1.png" width="66%" style="display: block; margin: auto;" />
]

.pull-right[

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", 
  y="bill_depth_mm",
  hue="species"
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-4-3.png" width="66%" style="display: block; margin: auto;" />
]

---

## A more complex plot

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  col="island", row="species"
)
```

---

## Figure-level vs. axes-level functions

.footnote[
These are not the only axes-level functions - we see additional plotting functions in a bit
]

---

## displots

.pull-left[

```python
sns.displot(
  data = penguins,
  x = "bill_length_mm", hue = "species",
  alpha = 0.5, aspect = 1.5
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-7-1.png" width="75%" style="display: block; margin: auto;" />
]

.pull-right[

```python
sns.displot(
  data = penguins,
  x = "bill_length_mm", hue = "species",
  kind = "kde", fill=True,
  alpha = 0.5, aspect = 1
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-8-3.png" width="75%" style="display: block; margin: auto;" />
]

---

## catplots

.pull-left[

```python
sns.catplot(
  data = penguins,
  x = "species", y = "bill_length_mm",
  hue = "sex"
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-9-5.png" width="75%" style="display: block; margin: auto;" />
]

.pull-right[

```python
sns.catplot(
  data = penguins,
  x = "species", y = "bill_length_mm",
  hue = "sex",
  kind = "box"
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-10-7.png" width="75%" style="display: block; margin: auto;" />
]

---

## figure-level plot size

To adjust the size of plots generated via a figure-level plotting function adjust the `aspect` and `height` arguments, figure width is `aspect * height`.

.pull-left[

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  aspect = 1, height = 3
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-11-9.png" width="66%" style="display: block; margin: auto;" />
]

.pull-right[

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  aspect = 1, height = 5
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-12-11.png" width="66%" style="display: block; margin: auto;" />
]

.footnote[
Note this is the size of a facet (Axes) not the figure
]

---

## figure-level plot details

.pull-left[

```python
g = sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  aspect = 1
)

g
```

```python
print(g)
```

```
## <seaborn.axisgrid.FacetGrid object at 0x294d757f0>
```
]

.pull-right[

```python
h = sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species", col="island",
  aspect = 1/2
)

h
```

```python
print(h)
```

```
## <seaborn.axisgrid.FacetGrid object at 0x28d4474f0>
```
]

---

## FacetGird methods

.small[
| Method              | Description     |
|---------------------|:-------------------------------------------------------------------------|
| `add_legend()`      | Draw a legend, maybe placing it outside axes and resizing the figure     |
| `despine()`	        | Remove axis spines from the facets.                                      |
| `facet_axis()`    	| Make the axis identified by these indices active and return it.          |
| `facet_data()`	    | Generator for name indices and data subsets for each facet.              |
| `map()`	            | Apply a plotting function to each facet’s subset of the data.            |
| `map_dataframe()`	  | Like `.map()` but passes args as strings and inserts data in kwargs.     |
| `refline()`	        | Add a reference line(s) to each facet.                                   |
| `savefig()`	        | Save an image of the plot.                                               |
| `set()`	            | Set attributes on each subplot Axes.                                     |
| `set_axis_labels()`	| Set axis labels on the left column and bottom row of the grid.           |
| `set_titles()`	    | Draw titles either above each facet or on the grid margins.              |
| `set_xlabels()`	    | Label the x axis on the bottom row of the grid.                          |
| `set_xticklabels()`	| Set x axis tick labels of the grid.                                      |
| `set_ylabels()`	    | Label the y axis on the left column of the grid.                         |
| `set_yticklabels()`	| Set y axis tick labels on the left column of the grid.                   |
| `tight_layout()`	  | Call fig.tight_layout within rect that exclude the legend.               |
]

---

## Adjusting labels

.pull-left[

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  aspect = 1
).set_axis_labels(
  "Bill Length (mm)", "Bill Depth (mm)"
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-15-17.png" width="66%" style="display: block; margin: auto;" />
]

.pull-right[

```python
sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species", col="island",
  aspect = 1/2
).set_axis_labels(
  "Bill Length (mm)", "Bill Depth (mm)"
).set_titles(
  "{col_var} - {col_name}" 
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-16-19.png" width="90%" style="display: block; margin: auto;" />
]

---

## FacetGrid attributes

| Attribute   | Description    |
|-------------|:--------------------------------------------------------------------|
| `ax`	      | The `matplotlib.axes.Axes` when no faceting variables are assigned. | 
| `axes`	    | An array of the `matplotlib.axes.Axes` objects in the grid.         | 
| `axes_dict` | A mapping of facet names to corresponding `matplotlib.axes.Axes`.   | 
| `figure`	  | Access the `matplotlib.figure.Figure` object underlying the grid.   | 
| `legend`    |	The `matplotlib.legend.Legend` object, if present.                  |

---

## Using axes to modify plots

.pull-left[

```python
g = sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  aspect = 1
)
g.ax.axvline(
  x = penguins.bill_length_mm.mean(), c="k"
)
```

<img src="Lec11_files/figure-html/unnamed-chunk-17-21.png" width="66%" style="display: block; margin: auto;" />
]

.pull-right[

```python
h = sns.relplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species", col="island",
  aspect = 1/2
)
mean_bill_dep = penguins.bill_depth_mm.mean()

[ ax.axhline(y=mean_bill_dep, c="c") 
  for row in h.axes for ax in row ]
```

<img src="Lec11_files/figure-html/unnamed-chunk-18-23.png" width="90%" style="display: block; margin: auto;" />
]

---

## Why figure-level functions

.pull-left[
#### Advantages:
* Easy faceting by data variables
* Legend outside of plot by default
* Easy figure-level customization
* Different figure size parameterization
]

.pull-right[
#### Disadvantages:
* Many parameters not in function signature
* Cannot be part of a larger matplotlib figure
* Different API from matplotlib
* Different figure size parameterization
]

.footnote[Details based on [seaborn docs](https://seaborn.pydata.org/tutorial/function_overview.html#relative-merits-of-figure-level-functions)]

---

## lmplots

There is one last figure-level plot type - `lmplot()` which is a convenient interface to fitting and ploting regression models across subsets of data,

```python
sns.lmplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species", col="island",
  aspect = 1, truncate=False
)
```

---

## axes-level functions

These functions return a `matplotlib.pyplot.Axes` object instead of a `FacetGrid` giving more direct control over the plot using basic matplotlib tools.

.pull-left[

```python
plt.figure()

sns.scatterplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species"
)

plt.xlabel("Bill Length (mm)")
plt.ylabel("Bill Depth (mm)")
plt.title("Length vs. Depth")

plt.show()
```
]

.pull-right[
<img src="Lec11_files/figure-html/unnamed-chunk-20-27.png" width="85%" style="display: block; margin: auto;" />
]

---

## subplots - pyplot style

.pull-left[

```python
plt.figure(figsize=(4,6), layout="constrained")

plt.subplot(211)
sns.scatterplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species"
)

plt.subplot(212)
sns.countplot(
  data=penguins,
  x="species"
)

plt.show()
```
]

.pull-right[
<img src="Lec11_files/figure-html/unnamed-chunk-21-29.png" width="66%" style="display: block; margin: auto;" />
]

---

## subplots - OO style

.pull-left[

```python
fig, axs = plt.subplots(
  2, 1, figsize=(4,6), 
  layout="constrained",
  sharex=True  
)

sns.scatterplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species",
  ax = axs[0]
)

sns.kdeplot(
  data=penguins,
  x="bill_length_mm", hue="species",
  fill=True, alpha=0.5,
  ax = axs[1]
)

plt.show()
```
]

.pull-right[
<img src="Lec11_files/figure-html/unnamed-chunk-22-31.png" width="66%" style="display: block; margin: auto;" />
]

---

## layering plots

.pull-left[

```python
plt.figure(layout="constrained")

sns.kdeplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species"
)

sns.scatterplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species", alpha=0.5
)

sns.rugplot(
  data=penguins,
  x="bill_length_mm", y="bill_depth_mm",
  hue="species"
)

plt.legend()

plt.show()
```
]

.pull-right[
<img src="Lec11_files/figure-html/unnamed-chunk-23-33.png" width="85%" style="display: block; margin: auto;" />
]

---

## Themes

Seaborn comes with a number of themes (`darkgrid`, `whitegrid`, `dark`, `white`, and `ticks`) which can be enabled by `sns.set_theme()` at the figure level or `sns.axes_style()` at the axes level.

.pull-left[ .small[

```python
def sinplot():
    x = np.linspace(0, 14, 100)
    for i in range(1, 7):
        plt.plot(x, np.sin(x + i * .5) * (7 - i))
        
sinplot()
plt.show()
```

<img src="Lec11_files/figure-html/unnamed-chunk-24-35.png" width="45%" style="display: block; margin: auto;" />
] ]

.pull-right[ .small[

```python
with sns.axes_style("darkgrid"):
  sinplot()
  plt.show()
```

<img src="Lec11_files/figure-html/unnamed-chunk-25-37.png" width="45%" style="display: block; margin: auto;" />
] ]

---

.pull-left[ .small[

```python
with sns.axes_style("whitegrid"):
  sinplot()
  plt.show()
  
```

```python
with sns.axes_style("dark"):
  sinplot()
  plt.show()
```

<img src="Lec11_files/figure-html/unnamed-chunk-26-40.png" width="45%" style="display: block; margin: auto;" />
] ]

.pull-right[ .small[

```python
with sns.axes_style("white"):
  sinplot()
  plt.show()
```

```python
with sns.axes_style("ticks"):
  sinplot()
  plt.show()
```

<img src="Lec11_files/figure-html/unnamed-chunk-27-44.png" width="45%" style="display: block; margin: auto;" />
] ]

---

## Context

.pull-left[ .small[

```python
sns.set_context("notebook")
sinplot()
plt.show()
  
```

```python
sns.set_context("paper")
sinplot()
plt.show()
```

<img src="Lec11_files/figure-html/unnamed-chunk-28-48.png" width="40%" style="display: block; margin: auto;" />
] ]

.pull-right[ .small[

```python
sns.set_context("talk")
sinplot()
plt.show()
```

```python
sns.set_context("poster")
sinplot()
plt.show()
```

<img src="Lec11_files/figure-html/unnamed-chunk-29-52.png" width="40%" style="display: block; margin: auto;" />
] ]

---

## Color palettes

All of the examples below are the result of calls to `sns.color_palette()` with `as_cmap=True` for the continuous case,

.pull-left[ .small[

```python
show_palette()
```

```python
show_palette("tab10")
```

```python
show_palette("hls")
```

```python
show_palette("husl")
```

```python
show_palette("Set2")
```

<img src="Lec11_files/figure-html/unnamed-chunk-32-59.png" width="75%" style="display: block; margin: auto;" />
] ]

.pull-right[ .small[

```python
show_cont_palette("cubehelix")
```

```python
show_cont_palette("light:b")
```

```python
show_cont_palette("dark:salmon_r")
```

```python
show_cont_palette("YlOrBr")
```

```python
show_cont_palette("vlag")
```

<img src="Lec11_files/figure-html/unnamed-chunk-33-69.png" width="75%" style="display: block; margin: auto;" />
] ]

.footnote[
See more examples in the color palettes [tutorial](https://seaborn.pydata.org/tutorial/color_palettes.html)
]

---

## Pair plots

```python
sns.pairplot(data = penguins, height=5)
```

---

```python
sns.pairplot(data = penguins, hue="species", height=5, corner=True)
```

---

## PairGrid

`pairplot()` is a special case of the more general `PairGrid` - once constructed there are methods that allow for mapping plot functions of the different axes,

```python
sns.PairGrid(penguins, hue="species", height=5)
```

---

## Mapping

.pull-left-narrow[

```python
g = sns.PairGrid(
  penguins, hue="species",
  height=3
)

g = g.map_diag(
  sns.histplot, alpha=0.5
)

g = g.map_lower(
  sns.scatterplot
)

g = g.map_upper(
  sns.kdeplot
)

g
```
]

.pull-right-wide[
<img src="Lec11_files/figure-html/unnamed-chunk-37-81.png" width="75%" style="display: block; margin: auto;" />
]

---

## Pair subsets

```python
x_vars = ["body_mass_g", "bill_length_mm", "bill_depth_mm", "flipper_length_mm"]
y_vars = ["body_mass_g"]

g = sns.PairGrid(penguins, hue="species", x_vars=x_vars, y_vars=y_vars, height=3)

g = g.map_diag(sns.kdeplot, fill=True)
g = g.map_offdiag(sns.scatterplot, size=penguins["body_mass_g"])
g = g.add_legend()

g
```

---

## Custom FacetGrids

Just like `PairGrid`s it is possible to construct `FacetGrid`s from scratch,

```python
sns.FacetGrid(penguins, col="island", row="species")
```

---

```python
g = sns.FacetGrid(penguins, col="island", hue="species")

g = g.map(sns.scatterplot, "bill_length_mm", "bill_depth_mm")
g = g.add_legend()

g
```

---

## Custom plots / functions

```python
from scipy import stats
def quantile_plot(x, **kwargs):
    quantiles, xr = stats.probplot(x, fit=False)
    plt.scatter(xr, quantiles, **kwargs)

g = sns.FacetGrid(penguins, col="species", height=3, sharex=False)
g.map(quantile_plot, "body_mass_g", s=2, alpha=0.5)
```

.footnote[Example from axis grid [tutorial](https://seaborn.pydata.org/tutorial/axis_grids.html#using-custom-functions)]

---

## jointplot

One final figure-level plot, is a joint plot which includes marginal distributions along the x and y-axis.

```python
g = sns.jointplot(data=penguins, x="bill_length_mm", y="bill_depth_mm", hue="species")
```

```
## /opt/homebrew/lib/python3.9/site-packages/seaborn/axisgrid.py:1740: UserWarning: Tight layout not applied. tight_layout cannot make axes width small enough to accommodate all axes decorations
##   f.tight_layout()
```

```python
plt.show()
```

---

## Adjusting

The main plot (joint) and the margins (marginal) can be modified by keywords or via layering (use `plot_joint()` and `plot_marginals()` methods).

```python
g = sns.jointplot(data=penguins, x="bill_length_mm", y="bill_depth_mm", hue="species", marginal_kws=dict(fill=False))
```

```python
g = g.plot_joint(sns.kdeplot, alpha=0.5, levels=5)
plt.show()
```