filter.Rmd

## Filtering Observations

## Filtering Observations

Recall this bit of code?

```{r bballRecall, eval=FALSE}
bball = read_html(url) %>% 
  html_nodes("#totals_stats") %>% 
  html_table %>% 
  data.frame %>% 
  filter(Rk != "Rk")
```

You will notice the <span class="func">filter</span> line at the end

We sometimes want to see a very specific portion of the data


## Filtering Observations

>- <span class='func'>filter</span> returns rows with matching conditions.
>- <span class='func'>slice</span> allows for a numeric indexing approach.

>- Say we want too look at forwards (SF or PF) over the age of 35...

```{r filter1, eval=FALSE}
bball %>% 
  filter(Age > 35, Pos == "SF" | Pos == "PF")
```

>- or the first 10 rows...

```{r filter2, eval=FALSE}
bball %>% 
  slice(1:10)
```


## Filtering Observations

This can be done with things that are created on the fly...

```{r uniteFilterArrange, cache=TRUE}
bball %>% 
  unite("posTeam", Pos, Tm) %>%             # create a new variable
  filter(posTeam == "PF_SAS") %>%           # use it for filtering
  arrange(desc(PTS/PF)) %>%                 # order by on-the-fly operation
  select(1:10)
```


## Your turn
## Your turn
A brief exercise:

- $1.$ <span class='func'>filter</span> the iris data set to only the `r "virginica"` species (var <span class='func'>==</span> value)
- $2.$ show only **Petal.Length** and **Petal.Width** variables (<span class='func'>select</span>)
- $3.$ <span class="emph">bonus</span>: redo, but instead, filter if the ratio of **Petal.Length** to **Petal.Width** is greater than 5. 
    
- Which species do these observations belong to?


## Example

```{r yourTurnFilterExample, cache=TRUE}
iris %>% 
  filter(Petal.Length/Petal.Width > 5) %>% 
  summary
```