---
title: 'ExercisesRmd'
output: html_document
date: '2023-04-11'
---

# Load data

Load package

```{r}
library(palmerpenguins)
```

Load data frame
```{r}
penguins<-penguins
```

## rows
Check first 5 rows
```{r}
head(penguins,5)
```

Check last 5 rows
```{r}
tail(penguins,5)
```

When you want to inspect specific rows, they are in the **first** position
```{r}
(penguins[1,])
```

Check first 3 rows
The **:** means "from A to B".
```{r}
(penguins[1:3,])
```

## columns
The columns go on the **second** position.  
```{r}
head(penguins[,1])
```

Another way to do it is with the $ and the column name.  
```{r}
head(penguins$species)
```

## column and row
```{r}
(penguins[1,1])
```

```{r}
(penguins[3,2])
```

# Tidyverse

Load package
```{r}
library(tidyverse)
```

## count
Sample size?
```{r}
penguins %>% 
  count()
```

Sample size per species?
```{r}
penguins %>% 
  count(species)
```

Per island and per species?
```{r}
penguins %>% 
  count(island,species)
```

## distinct
Using base R
```{r}
unique(penguins$species)
```

Using tidyverse
```{r}
penguins %>% 
  distinct(species)  
```

## select
Select one column
```{r}
penguins %>% 
  select(species) 
```

Remove one column using the **-** 
```{r}
penguins %>% 
  select(-sex) 
```

Select all columns expect this one using **!** 
```{r}
penguins %>% 
  select(!sex) 
```

Select columns in between using **:** 
```{r}
penguins %>% 
  select(bill_length_mm:body_mass_g) 
```

Select columns based on characters (strings).
```{r}
penguins %>% 
  select(ends_with("mm"))
```

Using the first letters of the string
```{r}
penguins %>% 
  select(starts_with("bill"))
```

## filter
```{r, eval=FALSE}
penguins %>%
  filter(sex == 'female')
```

To change the object we need to create a new data frame.
```{r}
female_penguins<-penguins %>%
  filter(sex == 'female')
```

- The symbol **<=** means 'smaller or same as'
```{r, eval=FALSE}
penguins %>%
  filter(bill_length_mm <= 39.1)  
```


- The symbol **>=** means 'larger or same as'
```{r, eval=FALSE}
penguins %>%
  filter(bill_length_mm >= 39.1) 
```

- The symbol **&** means 'and'
```{r, eval=FALSE}
penguins %>%
  filter(island == 'Biscoe' & species =='Adelie') 
```

## mutate
```{r}
penguins<-penguins %>% 
  mutate(body_mass_kg = body_mass_g / 1000) 
```

## group_by
```{r}
penguins %>% 
  group_by(year) %>%
  summarise(mean_bill_length=mean(bill_length_mm))
```

## drop_na 

This functions allows you to ignore or remove NAs
```{r, eval=FALSE}
penguins %>% 
  drop_na(bill_length_mm)
```

## group_by
```{r}
penguins %>% 
  group_by(year) %>% 
  drop_na(bill_length_mm) %>%
  summarise(mean_bill_length=mean(bill_length_mm))
```

### fill out

- Create a new data frame with different penguin species 
```{r}
#______ <-penguins %>%
#  distinct(_____)
```

- Create a new object with only two columns: sex and year
```{r}
#_____ <-penguins %>%
#  select(____,_____)
```

- Create a new data frame only with male data
```{r}
#______ <-penguins %>%
#  filter(_____ == '______')
```

- Create a new data frame using only females that are heavier than 3800 g
```{r}
#______ <-penguins %>%
#  filter(_____ == '______')%>%
#  filter(______ >= ______)
```

- Check the results form this operation
```{r}
penguins_means<-penguins %>% 
  group_by(species,sex) %>% 
  drop_na(body_mass_g,sex)%>% 
  summarise(mean_body_mass_g = mean(body_mass_g), n = n())%>% 
  mutate(mean_body_mass_kg = mean_body_mass_g / 1000)
```
# Tables

```{r}
bird_id<-c("ID01","ID02","ID03","ID04","ID05",
      "ID06","ID07","ID08","ID09","ID10")
bird_mass<-c(1.5,2.0,3.5,4.1,2.6,3.7,8.9,2.5,6.3,1.0)
bird_gps<-c(50010,50020,50035,50001,50006,50003,50008,50002,50003,50001)
```

We might have two data sets  
On one hand, the measurement data...
```{r}
bird_measurements<-data.frame(bird_id,bird_mass)
head(bird_measurements)
```

... on the other, field data.
```{r}
bird_tracking <- data.frame(bird_id,bird_gps)
head(bird_tracking)
```

## left_join()

To join them we can use the function **left_join()**  
However it is important to have a **key** to match the observations

```{r}
bird_joined<-left_join(bird_measurements,
                       bird_tracking, 
                       by = "bird_id")  
bird_joined
```



## pivot_longer

Lets imagine we have data from five species and their number of locations (nlocs) among three different years. 

```{r}
bird_id<-c("ID01","ID02","ID03","ID04","ID05",
      "ID06","ID07","ID08","ID09","ID10")
year_2010<-c(5,4,5,6,7,3,2,1,9,10)
year_2011<-c(3,2,1,9,4,5,6,7,3,2)
year_2012<-c(6,2,3,7,8,2,1,9,4,5)
```

New data frame
```{r}
bird_nlocs<-data.frame(bird_id,year_2010,year_2011,year_2012)
```

```{r}
head(bird_nlocs,5)
```

## pivot_longer

pivot_longer "lengthens" data, increasing the number of rows and decreasing the number of columns. 

```{r}
bird_long <- bird_nlocs %>%
  pivot_longer(c(year_2010,year_2011,year_2012),
    names_to = "year",
    values_to = "nlocs" )
```

```{r}
head(bird_long,5)
```

## pivot_wider

The opposite will be to separate the columns.

pivot_wider() "widens" data, increasing the number of columns and decreasing the number of rows.

The most important arguments are **names_from** which are going to be the names of the columns created after (often the column with factors) and **values_from** is the the name of the column with the values (often the columns with numbers)

```{r}
bird_wide<-bird_long %>% 
  pivot_wider(names_from = year, 
              values_from = nlocs)
bird_wide
```

## paste or unite

The argument **paste** or **paste0** from base R allows you to paste together multiple columns

```{r}
bird_long$unique_id<-paste0(bird_long$bird_id,'_',bird_long$year)
```

The argument **unite** is similar, but lets you to paste together multiple columns into one.

```{r}
bird_long<-bird_long %>% 
  unite(col = unique_id2,  
        c("bird_id", "year"),
        sep = "_",
        remove=FALSE)
bird_long
```

**Note** it will get rid of the original column, so if you don't want to eliminate the original column add **remove = FALSE.**

## separate

```{r, eval=FALSE}
bird_long %>% 
  separate(col = unique_id,  
           into = c("id", "text","year"),
           sep = "_")
```

```{r,}
head(bird_long,5)
```

**Note** it will get rid of the original column, so if you don't want to eliminate the original column add **remove = FALSE.**

```{r}
bird_long<-bird_long %>% 
  separate(col = unique_id,  
           into = c("id", "text","year"),
           sep = "_",
           remove = FALSE)  
```

## rename

The argument **rename** allows to change the name of one or several columns.  
The new name is written first and the old name comes after.

An example changing the name of one column

```{r, eval=FALSE}
bird_long %>% 
  rename(unique_identifier = unique_id2)  
```

## relocate

The argument **relocate** allows you to reorganize your columns and keeping just those that you are interested on.

```{r, eval=FALSE}
bird_long %>% 
  relocate(bird_id,year,nlocs)  
```

Using this argument together with select you can keep only the columns of interest.

```{r, eval=FALSE}
bird_long %>% 
  select(bird_id,year,nlocs)%>%
  relocate(bird_id,year,nlocs)  
```

# Export

```{r, message=FALSE}
library(here)
ResultsFolder<-here()
```

```{r, eval=FALSE}
write_csv(
  bird_joined,
  file =paste0(ResultsFolder,'/bird_joined.csv'))
```