05-data-viz.qmd

# Data Exploration and Visualization

## Stacked ggplot with custom periodization colors

How to create a stacked bar chart to see your gps for the week. Label with your morphocycle/tactical periodization labels. And as a kicker - start your week on Thursday since that is your first game. Final output and code will look like this:

```{r final-first}
#| code-fold: true
#| warning: false
library(lubridate)
library(tibble)
library(dplyr)
library(ggplot2)
library(gt)

# DATA -----------------------------------------------------------------------------------------------------------------
# Create our own data
raw_gps_df <- tibble(
  date = seq.Date(from = as.Date('2023-08-01'), to = as.Date('2023-08-14'), 
                  by = 'days'),
  metric = c(0, 8, 12, 18, 6, 12, 20, 0, 6, 9, 7, 19, 4, 17)
  )
  
raw_session_desc <- tibble(
  date = seq.Date(from = as.Date('2023-08-01'), to = as.Date('2023-08-14'), 
                  by = 'days'),
  tp_code = c('Off', 'ReEntry', 'Strength', 'Match 1', 'Recovery', 'Activation',
              'Match 2', 'Off', 'Strength', 'Speed', 'Activation', 'Match 1',
              'ReEntry-Activation', 'Match 2')
  )

# Color codes. Presented a little cleaner to read (for us humans)
raw_tp_codes <-tribble(
~tp_code, ~tp_color,
'Off', 'grey',
'ReEntry', 'lightgreen',
'Strength', 'blue',
'Endurance', 'limegreen',
'Speed', 'yellow',
'Activation', '#fff5a9',
'ReEntry-Activation','lightgreen',
'Match 1', 'darkgreen',
'Match 2', 'darkgreen',
'Recovery', 'chartreuse'
)

# CUSTOM FUNCTION ------------------------------------------------------------------------------------------------------

# Custom function to change a day ("Sunday", "Monday", "Tuesday", etc) 
# into a number with Sun = 1. Surprised this isn't a function in lubridate.
daynum <- function(day){
  
  day <- tolower(day)
  
  if(grepl("su", day)) {
    return(1)
    
  } else if (grepl("mo", day)) {
    return(2)
    
  } else if (grepl("tu", day)) {
    return(3)
  
  } else if (grepl("we", day)) {
    return(4)
    
  } else if (grepl("th", day)) {
    return(5)
    
  } else if (grepl("fr", day)) {
    return(6)

  } else if (grepl("sa", day)) {
    return(7)
    
  } else {
    return(NA)
  }
}

# This is the order we want our week to be in. 
# Thursday is the first day, Wednesday is the last.

week_order <- c('Thu', 'Fri', 'Sat', 'Sun', 'Mon', 'Tue', 'Wed' )

# TRANSFORM ------------------------------------------------------------------------------------------------------------

# Combine into one data frame
session_df <- raw_gps_df |>
  left_join(raw_session_desc, by = 'date') |>
  left_join(raw_tp_codes, by = 'tp_code') |>
  mutate(week = floor_date(date, unit = 'weeks', daynum(week_order[1])))

cols <- distinct(session_df, tp_code, tp_color) |> deframe()

# PLOT -----------------------------------------------------------------------------------------------------------------

session_df |>
  ggplot(aes(x = week, y = metric, fill = tp_code, group = date)) +
  geom_col(position = position_stack(reverse = TRUE)) +
  scale_fill_manual(
    name = "Tact Periodization Days",
    values = cols,
  ) +
  scale_x_date(
    breaks = session_df$week,
    date_labels = "%m-%d") +
  theme_minimal() 

```

First lets create some fake data to use. You would probably import these with `read_csv`, but I'm going to keep it simple and just create it here.

```{r}
#| warning: false
library(lubridate)
library(tibble)
library(dplyr)
library(ggplot2)
library(gt)

# Create our own data
raw_gps_df <- tibble(
  date = seq.Date(from = as.Date('2023-08-01'), to = as.Date('2023-08-14'), 
                  by = 'days'),
  metric = c(0, 8, 12, 18, 6, 12, 20, 0, 6, 9, 7, 19, 4, 17)
  )
  
raw_session_desc <- tibble(
  date = seq.Date(from = as.Date('2023-08-01'), to = as.Date('2023-08-14'), 
                  by = 'days'),
  tp_code = c('Off', 'ReEntry', 'Strength', 'Match 1', 'Recovery', 'Activation',
              'Match 2', 'Off', 'Strength', 'Speed', 'Activation', 'Match 1',
              'ReEntry-Activation', 'Match 2')
  )

# Look at the data
raw_gps_df |> gt()
raw_session_desc |> gt()
```

I usually import my data as `raw_data <- ...` as it's a pain to trouble shoot and have to and have to to rerun the import from the top of the code.

```{r}
# Combine data
session_df <- raw_gps_df |>
  left_join(raw_session_desc, by = 'date')

# View combined data
session_df |> head() |> gt()

```

```{r}
# Plot data
session_df |>
  ggplot(aes(x = date, y = metric, fill = tp_code)) +
  geom_col() +
  theme_minimal()

```

It would be nice to use our own custom color scheme. Help on using custom color and labels in ggplot found [here](https://stackoverflow.com/questions/68557812/use-custom-color-and-custom-label-from-dataframe-columns-in-ggplot2). The 'trick' is to use `deframe`, which converts a two-column data frames to a named vector or list, using the first column as name and the second column as value. If the input has only one column, an unnamed vector is returned

```{r}
# Color codes. Presented a little cleaner to read (for us humans)
raw_tp_codes <-tribble(
~tp_code, ~tp_color,
'Off', 'grey',
'ReEntry', 'lightgreen',
'Strength', 'blue',
'Endurance', 'limegreen',
'Speed', 'yellow',
'Activation', '#fff5a9',
'ReEntry-Activation','lightgreen',
'Match 1', 'darkgreen',
'Match 2', 'darkgreen',
'Recovery', 'chartreuse'
)


# Import and combine data
session_df <- raw_gps_df |>
left_join(raw_session_desc, by = 'date')


# deframe() converts two-column data frames to a named vector or list
cols <- distinct(raw_tp_codes, tp_code, tp_color) |> deframe()

# Following not used but may come in handy in the future
# labs \<- distinct(raw_tp_codes, tp_code, tp_code) |> deframe()

# PLot
session_df |>
ggplot(aes(x = date, y = metric, fill = tp_code)) +
geom_col(position='stack', stat='identity') +
scale_color_manual(aesthetics = 'fill',
values = cols
) +
theme_minimal()
```

```{r}

# Custom function to change a day ("Sunday", "Monday", "Tuesday", etc) 
# into a number with Sun = 1. Surprised this isn't a function in lubridate.
daynum <- function(day){
  
  day <- tolower(day)
  
  if(grepl("su", day)) {
    return(1)
    
  } else if (grepl("mo", day)) {
    return(2)
    
  } else if (grepl("tu", day)) {
    return(3)
  
  } else if (grepl("we", day)) {
    return(4)
    
  } else if (grepl("th", day)) {
    return(5)
    
  } else if (grepl("fr", day)) {
    return(6)

  } else if (grepl("sa", day)) {
    return(7)
    
  } else {
    return(NA)
  }
}

# This is the order we want our week to be in. 
# Thursday is the first day, Wednesday is the last.
week_order <- c('Thu', 'Fri', 'Sat', 'Sun', 'Mon', 'Tue', 'Wed' )

# Combine 
session_df <- raw_gps_df |>
  left_join(raw_session_desc, by = 'date') |>
  left_join(raw_tp_codes, by = 'tp_code') |>
  mutate(week = floor_date(date, unit = 'weeks', daynum(week_order[1])))

cols <- distinct(session_df, tp_code, tp_color) |> deframe()

session_df |>
  ggplot(aes(x = week, y = metric, fill = tp_code, group = date)) +
  geom_col(position = position_stack(reverse = TRUE)) +
  scale_fill_manual(
    name = "Tact Periodization Days",
    values = cols,
  ) +
  scale_x_date(
    breaks = session_df$week,
    date_labels = "%m-%d") +
  theme_minimal() 

```