siuba-demo.Rmd

---
title: "5 Minutes to Siuba"
output:
  html_document:
    toc: true
    toc_float: true
classoption:
  - twocolumn
---

```{r}
#reticulate::use_virtualenv(path.expand("~/.virtualenvs/siuba"), required = TRUE)
reticulate::py_run_string("import matplotlib; matplotlib.use('Agg');")

knitr::opts_chunk$set(warning = FALSE, message = FALSE)
```

```{python, python-setup}
import pandas as pd
pd.set_option('display.max_rows', 5)

import matplotlib
matplotlib.use('Agg')
```



## What is siuba? {.tabset .tabset-pills}

### dplyr 

```{r}
library(dplyr)
library(datasets)

mtcars %>%
  group_by(cyl) %>%
  summarize(avg_hp = mean(hp))
```

### siuba

```{python}
from siuba import *
from siuba.data import mtcars

(mtcars
  >> group_by(_.cyl)
  >> summarize(avg_hp = _.hp.mean())
  )
```

### pandas

```{python}
import pandas as pd
from siuba.data import mtcars    # a pandas DataFrame


(mtcars
  .groupby("cyl", as_index = False)
  .agg(avg_hp = ("hp", "mean"))
  )
```

## {-}

## Why siuba?

### select columns

:::::: {.columns}

::: {.column width="50%"}

#### siuba

```{python, eval = FALSE}
# select two columns
select(mtcars, _.hp, _.mpg)

# drop two columns
select(mtcars, -_.hp, -_.mpg)

# rename a column
select(mtcars, _.horse_power == _.hp, _.mpg)



# match column names
select(mtcars, _.endswith("g"))


```

:::

::: {.column width="50%"}

#### pandas

```{python, eval = FALSE}
# select two columns
mtcars[["hp", "mpg"]]

# drop two columns
mtcars.drop(columns = ["hp", "mpg"])

# rename a column
(mtcars[["hp", "mpg"]]
  .rename(columns = {"hp": "horse_power"})
  )

# match column names
mtcars.loc[:, lambda x: 
      x.columns.str.endswith("g")
      ]

```

:::

::::::

<br>

### group by columns {.tabset .tabset-pills}

The code below demeans a column in pandas, but does not work with grouped data :/.

```{python, results = "hide"}
# mutate using hp
mtcars.eval("res = hp - hp.mean()")
```

Below shows how we could do it for grouped data.

#### siuba

```{python}
(mtcars
  >> group_by(_.cyl)
  >> mutate(res = _.hp - _.hp.mean())
  )
```

#### pandas

```{python}
(mtcars
  #.groupby("cyl") here would error
  .assign(res = lambda d: d.hp - d.groupby("cyl").hp.transform("mean"))
  )

```

### {-}

<br>

### SQL queries {.tabset .tabset-pills}

#### siuba

```{python}
# with DataFrame
(mtcars 
  >> group_by(_.cyl)
  >> summarize(avg_hp = _.hp.mean())
  )
```


```{python}
# with SQL
from siuba.data import cars_sql
from siuba import *

q = (cars_sql
  >> group_by(_.cyl)
  >> summarize(avg_hp = _.hp - _.hp.mean())
  >> show_query()
  )

```

#### pandas 

```{python}
# with DataFrame
(mtcars 
  .groupby("cyl", as_index = False)
  .agg(avg_hp = ("hp", "mean"))
  )
```


```{python, eval = FALSE}
# with SQL
pd.read_sql(engine, """
  SELECT cyl, AVG(hp) as avg_hp
  FROM mtcars
  GROUP BY cyl
""")

```

### {-}

<br>

### bonus ggplot

```{python}
import pandas as pd

from siuba import *
from plotnine import *     # ggplot for python

baltimore_bridges = pd.read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2018/2018-11-27/baltimore_bridges.csv")

(baltimore_bridges
  >> count(decade = (_.yr_built // 10) * 10)
  >> ggplot(aes("decade", "n"))
   + geom_line()
  )
```