In AndersAskeland/LearnR3: What the Package Does (One Line, Title Case)

1. Importing my data usring vroom

As I understand vroom is just an automated readr.

library(vroom)
library(here)
library(dplyr)
library(purrr)
library(fs)
library(tidyverse)

This is a example of how to do it. We first create a reference to the dir, look at the data using spec, and import is using vroom. We manually specify the different column types.

# Data dir
user_1_info_file <- here("data-raw/mmash/user_1/user_info.csv")

# Load data first
vroom(user_1_info_file) %>%
  spec()

# Read data
user_1_info_data <- vroom(user_1_info_file,
                          col_select = -1,
                          col_types = cols(
                            Gender = col_character(),
                            Weight = col_double(),
                            Height = col_double(),
                            Age = col_double()),
                          .name_repair = snakecase::to_snake_case)

user_1_info_data

Exercise: Import the saliva data

Same as above

# Data dir
user_1_saliva_file <- here("data-raw/mmash/user_1/saliva.csv")

# look at data structure
vroom(user_1_saliva_file) %>%
  spec()

# Create final data
user_1_saliva_data <- vroom(user_1_saliva_file,
                                 col_select = -1,
                                 col_types = cols(
                                   SAMPLES = col_character(),
                                   `Cortisol NORM` = col_double(),
                                   `Melatonin NORM` = col_double()),
                            .name_repair = snakecase::to_snake_case)
user_1_saliva_data

2 - Import larger datasets

If you have a very big dataset, a good idea could be to read in a few lines (100) using spec() and look at this data prior to attempting to import a larger dataset.

# User dir
user_1_rr_file <- here(("data-raw/mmash/user_1/RR.csv"))

# Find column types
vroom(user_1_rr_file, n_max = 100) %>%
  spec()

# Import data
user_1_rr_data <- vroom(file = user_1_rr_file,
                        col_select = -1,
                        col_types = 
                          cols(
                              ibi_s = col_double(),
                              day = col_double(),
                              time = col_time(format = "")
                              ),
                        .name_repair = snakecase::to_snake_case)
user_1_rr_data

Exercise: Import the Actigraph data.

# User dir
user_1_Actigraph_file <- here(("data-raw/mmash/user_1/Actigraph.csv"))

# Find column types
vroom(user_1_Actigraph_file, n_max = 100) %>%
  spec()

# Import data
user_1_Actigraph_data <- vroom(file = user_1_Actigraph_file,
                        col_select = -1,
                        col_types = 
                          cols(
  Axis1 = col_double(),
  Axis2 = col_double(),
  Axis3 = col_double(),
  Steps = col_double(),
  HR = col_double(),
  `Inclinometer Off` = col_double(),
  `Inclinometer Standing` = col_double(),
  `Inclinometer Sitting` = col_double(),
  `Inclinometer Lying` = col_double(),
  `Vector Magnitude` = col_double(),
  day = col_double(),
  time = col_time(format = "")
                          ),
                        .name_repair = snakecase::to_snake_case)
user_1_Actigraph_data

3 - Functions

The functions are pasted tp the function R script.

My loops are bad. So we are going to use map to create a loop instead.

4 - Load data

Data is loaded manually

# List (create a vector) of all files that containe (regexpr) specific string in a specified folder.
user_info_files <- dir_ls(here("data-raw/mmash/"), 
                          regexp = "user_info.csv", 
                          recurse = TRUE)
# Function
user_info_df <- map_dfr(user_info_files, import_user_info, .id = "file_path_id") %>%
  extract_user_id()

Data is loaded using a function

# I have not made a function for this.

Exercise: Map on the other datasets

A better way, where it the columns are controlled for. Is based on function extract_data which check the data

Here the datasets are imported

saliva_df <- import_multiple_files_new("saliva")
rr_df <- import_multiple_files_new("RR")
actigraph_df <- import_multiple_files_new("Actigraph")

Using regex for user ID

# Way one
user_info_df %>% 
  mutate(user_id = str_extract(file_path_id, "user_[:digit:][:digit:]?"), .keep = "unused")

# Alternative
user_info_df %>% 
  mutate(user_id = str_extract(file_path_id, "user_[:digit:][:digit:]?")) %>% 
  select(-file_path_id)

# Do using function
extract_user_id(df = user_info_df)

Join datasets

full_join(user_info_df, saliva_df, by = "user_id")

We can add multiple columns using reduce

Full code first <- add_numbers(1,2) second <- add_numbers(first,3) third <- add_numbers(second,4) forth <- add_numbers(thir, 5)

Using reduce reduce(1:4, add_numbers)

combined_data <- reduce(list(user_info_df, saliva_df, actigraph_df, rr_df), full_join, by = "user_id")

# Only age
user_info_df %>% 
  select(age)

# All characters
user_info_df %>% 
  select(where(is.character))

# Summarize stuff
saliva_df %>% summarise(cortisol_mean = mean(cortisol_norm))

saliva_df %>% 
  summarise(cortisol_mean = mean(cortisol_norm),
            melatonin_mean = mean(melatonin_norm))

# Use of across function
saliva_df %>%
  summarise(across(c(cortisol_norm, melatonin_norm), list(mean = mean), na.rm = T)) 

saliva_df %>% 
  summarise(across(where(is.numeric), list(mean, sd)))

# Group by
rr_df %>% 
  group_by(user_id, day) %>% # we group by two varibles
  summarise(across(ibi_s, list(missing = ~sum(is.na(.x)), mean = mean, sd = sd), na.rm = T), .groups = "drop_last")

# You can also drop the error with dplyr.summraise.inform
options(dplyr.summarise.inform = F)

Lets combine stuff

summarised_rr_df <- rr_df %>% 
  group_by(user_id, day) %>% # we group by two varibles
  summarise(across(ibi_s, list(mean = mean, sd = sd), na.rm = T), .groups = "drop_last")

Combine datasets

reduce(list(user_info_df, saliva_df, summarised_rr_df), full_join, by = "user_id")

user_info_df %>% mutate(age_category = case_when(age > 20 ~ "old",
                                                 age <= 20 ~ "young",
                                                 TRUE ~ NA_character_)) # Avoid mixing of characters with numeric)


saliva_with_day_df <- saliva_df %>% 
  mutate(day = case_when(
    samples == "before sleep" ~ 1,
    samples == "wake up" ~ 2,
    TRUE ~ NA_real_))

reduce(list(user_info_df, saliva_with_day_df, summarised_rr_df), full_join)

Exersise

# Summarise hr and steps
summarised_actigraph_df <- actigraph_df %>% 
  group_by(user_id, day) %>% # we group by two varibles
  summarise(across(c(hr, steps), list(mean = mean, sd = sd, max = max()), na.rm = T), .groups = "drop_last")
summarised_actigraph_df

# Combine data
reduce(list(actigraph_df, summarised_actigraph_df), full_join)

Exercise: Importing activitiy data

# Find and look at data
### Data dir
activity_dir <- here("data-raw/mmash/user_1/Activity.csv")

### Load data first
vroom(activity_dir) %>%
  spec()

# Extract data using function
activity_df <- import_multiple_files_new("Activity")

# Look at data
head(activity_df)
colnames(activity_df)

# Summarise hr and steps
activity_df <- activity_df %>% 
  mutate(activity_seconds = (end-start), .keep = "all")

activity_df %>% 
  group_by(activity, user_id)








summarised_actigraph_df <- actigraph_df %>% 
  group_by(user_id, day) %>% # we group by two varibles
  summarise(across(c(hr, steps), list(mean = mean, sd = sd, max = max()), na.rm = T), .groups = "drop_last")
summarised_actigraph_df

# Combine data
reduce(list(actigraph_df, summarised_actigraph_df), full_join)

activity_df$activity <-  factor(activity_df$activity, levels = c(1:12), labels = c("Low", "High", "Test", "4", "5", "6", "7", "8", "9", "10", "11", "12"))

)

# import data
user_info_df <- import_multiple_files_new("info")
saliva_df <- import_multiple_files_new("saliva")
rr_df <- import_multiple_files_new("RR")
actigraph_df <- import_multiple_files_new("Actigraph")

# Summarise RR data
summarised_rr_df <- rr_df %>% 
  group_by(user_id, day) %>% # we group by two varibles
  summarise(across(ibi_s, list(mean = mean, sd = sd), na.rm = T), .groups = "drop_last")

# Mutate saliva data

AndersAskeland/LearnR3 documentation built on Sept. 15, 2020, 10:39 a.m.