python-grammer-of-graphics.qmd

---
title: "Intro to Grammar of Graphics"
subtitle: "seaborn vs plotnine"
author: "Credit to @ChelseaParlett"
jupyter: venv-prez-kernel
format: 
    revealjs:
        height: 1400
        width: 1920
execute:
    echo: true
    eval: true
---

## 

<blockquote class="twitter-tweet"><p lang="en" dir="ltr">Started translating my 📊plotnine data viz lectures into 📈seaborn (so students can see both). <br><br>💭and I have some thoughts.<br><br>Here&#39;s the notebook with plotnine (commented out) and seaborn plots: <a href="https://t.co/Z36rodGFUU">https://t.co/Z36rodGFUU</a> <a href="https://t.co/Hw39RxbVS9">pic.twitter.com/Hw39RxbVS9</a></p>&mdash; Chelsea Parlett-Pelleriti (@ChelseaParlett) <a href="https://twitter.com/ChelseaParlett/status/1557482309565177859?ref_src=twsrc%5Etfw">August 10, 2022</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>

## Library and Prep

```{python}
import warnings
warnings.filterwarnings('ignore')

import pandas as pd
import numpy as np
from plotnine import *
# from plotnine.data import mtcars
import matplotlib.pyplot as plt
import seaborn as sns
mtcars = pd.read_csv("https://gist.githubusercontent.com/seankross/a412dfbd88b3db70b74b/raw/5f23f993cd87c283ce766e7ac6b329ee7cc2e1d1/mtcars.csv")


figure_size = (6.4, 4.8)
%matplotlib inline
```

## Scatterplot

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(mtcars, aes(x = "wt", y = "mpg")) + geom_point())
```


## seaborn

```{python}
sns.scatterplot(x = "wt", y = "mpg", data = mtcars)
```

:::

## Smooth

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(mtcars, aes(x = "wt", y = "mpg")) + 
    geom_point() +
    stat_smooth(method = "lm"))
```

## seaborn

```{python}
sns.regplot(x = "wt", y = "mpg", data = mtcars)
```

:::

## Facets

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(mtcars, aes(x = "wt", y = "mpg")) + geom_point() +
stat_smooth(method = "lm") + 
facet_wrap("~gear"))
```

## Seaborn

```{python}


facet = sns.FacetGrid(col = "gear", data = mtcars)
facet.map(sns.regplot, "wt", "mpg")
```

:::

## Smooth + Facet

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(mtcars, aes(x = "wt", y = "mpg", color = "factor(gear)")) + geom_point() +
stat_smooth(method = "lm") + 
facet_wrap("~gear"))
```


## Seaborn

```{python}


facet = sns.FacetGrid(col = "gear", data = mtcars, hue = "gear")
facet.map(sns.regplot, "wt", "mpg")
```

:::

## Bring on the penguins!

```{python}
penguin = pd.read_csv("https://raw.githubusercontent.com/cmparlettpelleriti/CPSC392ParlettPelleriti/master/Data/penguins.csv")
penguin.head()
```

# SNS Basic Steps

1. Tell SNS what kind of plot you want (using `sns.*`)

2. Tell SNS what you want to plot (inside the `sns.*()` function)

## Scatterplot

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "bill_length_mm", y = "bill_depth_mm", color = "species")) + geom_point())
```

## seaborn

```{python}
sns.scatterplot(x = "bill_length_mm", y = "bill_depth_mm", hue = "species", data = penguin)
```

:::

## Histogram

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "body_mass_g")) + geom_histogram())
```

## seaborn

```{python}
sns.histplot(x = "body_mass_g", data = penguin)
```

:::

## Bar

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "species")) + geom_bar())
```

## Seaborn

```{python}
sns.countplot(x = "species", data = penguin)
```

:::

## Boxplot

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "species", y = "bill_length_mm")) + geom_boxplot())
```

## seaborn

```{python}
sns.boxplot(x ="species", y = "bill_length_mm", data = penguin)
```

:::

## Boxplot + Point

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "species", y = "bill_length_mm", fill = "species")) +
  geom_point() + 
  geom_boxplot() + theme_minimal())
```


## seaborn

```{python}
sns.boxplot(x ="species", y = "bill_length_mm", data = penguin)
sns.scatterplot(x = "species", y = "bill_length_mm", data = penguin, color = "black")
```

:::

## Color by species

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "species", y = "bill_length_mm")) +
 geom_point() + 
 geom_boxplot(aes(fill = "species")) + theme_minimal())
```

## seaborn

```{python}
sns.set_style("white")
sns.boxplot(x ="species", y = "bill_length_mm", data = penguin)
sns.scatterplot(x = "species", y = "bill_length_mm", data = penguin, color = "black")
```

:::

## Prepping Data for Barcharts

```{python}
penguin_bill = penguin.groupby(["species"], as_index = False)["bill_length_mm"].mean()
penguin_bill
```

## Plotting bar charts

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(penguin, aes(x = "species", y = "body_mass_g", fill = "species")) + 
stat_summary(fun_data = "mean_sdl", geom = "bar"))
```

## seaborn

```{python}
sns.barplot(x = "species", y = "body_mass_g", hue = "species", data = penguin, dodge = False)
```

:::

## Better bar chart

::: {.panel-tabset}

## plotnine


```{python}
(ggplot(penguin, aes(x = "species", y = "body_mass_g", fill = "species")) + 
 stat_summary(fun_data = "mean_sdl", geom = "bar") + 
  labs(x = "Species", y = "Body Mass (g)") + 
  ggtitle("Penguin Body Mass by Species") +
  theme_minimal() +
  theme(panel_grid_major_x = element_blank(),
      panel_grid_minor_x = element_blank(),
       panel_grid_minor_y = element_blank(),
      legend_position = "none"))
```

## seaborn

```{python}
bp = sns.barplot(x = "species", y = "body_mass_g", hue = "species", data = penguin, dodge = False)
bp.set(title = "Penguin Body Mass by Species",
xlabel = "Species", ylabel = "Body Mass (g)")
bp.grid(False)
plt.legend([],[], frameon = False)
```
:::

## Better bar chart

::: {.panel-tabset}

## plotnine

```{python}
 (ggplot(penguin_bill, aes(x = "species", y = "bill_length_mm", fill = "species")) + 
 geom_bar(stat = "identity") + theme_minimal() + 
 labs(x = "Species", y = "Average Bill Length (mm)"))
```

## seaborn

```{python}
bp2 = sns.barplot(x = "species", y = "bill_length_mm", hue = "species", data = penguin_bill, dodge = False)
bp2.set(xlabel = "Species", ylabel = "Average Bill Length (mm)")
plt.legend([],[], frameon = False)
```

:::

# NON Count Bar Charts



```{python}
books = ["Home Before Dark", "The Wives", "You", "The Last Mrs. Parrish", "The Guest List", "Invisible Girl"]
ratings = [4.08,3.63,3.93,3.93, 3.85, 3.81]
book_df = pd.DataFrame({"books":books, "ratings": ratings})
book_df
```

## NON count bar charts

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(book_df, aes(x = "books", y = "ratings")) + 
 geom_bar(aes(fill = "books"), stat = "identity") +
 theme_minimal() +
 labs(title = "Book Ratings",
     x = "Book Titles",
     y = "Average GoodRead Ratings") +
 theme(axis_text_x = element_text(angle = 45),
     legend_position = "none"))
```

## seaborn

```{python}
bp3 = sns.barplot(data = book_df, x = "books", y = "ratings", hue = "books", dodge = False)
bp3.set(title = "Book Ratings",
xlabel = "Book Titles", ylabel = "Average GoodRead Ratings")
plt.xticks(rotation=45)
plt.legend([],[], frameon = False)
```

:::

# Let's Try to explore a new data set using Visualization!

```{python}
cereal = pd.read_csv("https://raw.githubusercontent.com/reisanar/datasets/master/Cereals.csv")
cereal.head()
```

## countplot

::: {.panel-tabset}

## plotnine

```{python}
(ggplot(cereal, aes(x = "mfr", fill = "mfr")) + geom_bar() + theme_bw())
```

## seaborn

```{python}
sns.countplot(data = cereal, x = "mfr", hue = "mfr", dodge = False)
```

:::

## Facet wrap/grid

::: {.panel-tabset}

## plotnine

```{python}
# facet wrap/grid
(ggplot(cereal, aes(x = "protein", y = "sugars")) + geom_point() + theme_bw() +
 facet_wrap("~mfr"))
```

## seaborn

```{python}
f = sns.FacetGrid(data = cereal, col = "mfr", col_wrap= 3)
f.map(sns.scatterplot, "protein", "sugars")
```

```{python}
# (ggplot(cereal, aes(x = "mfr", y = "sugars")) + geom_boxplot() + theme_minimal())

sns.boxplot(data = cereal, x = "mfr", y = "sugars")
```

:::