ActivityProfileR

Overview

ActivityProfileR is an implementation of the functional ANOVA test for activity data described in Nonparametric comparisons of activity profiles from wearable device data by Hsin-wen Chang and Professor Ian McKeague. A presentation on the original paper can be found here.

Installation

# install.packages("remotes")
remotes::install_github("thecbp/ActivityProfileR")

Usage

To use ActivityProfileR, the package expects activity data in a particular format. You can see an example of this data format using the prepare_sim_data() function. The prepare_sim_data() expects a list of lists, where each inner lists contains parameters for a group you want to appear in the data. Activity data is generated through an Ornstein-Uhlenbeck process (implemented in the ESGtoolkit package).

library(ActivityProfileR)

# Load data for 3 groups with different characteristics

data(example_activity)

head(example_activity)
## # A tibble: 6 x 3
##   subject_id group Xt           
##        <int> <dbl> <list>       
## 1          1     1 <dbl [1,000]>
## 2          2     1 <dbl [1,000]>
## 3          3     1 <dbl [1,000]>
## 4          4     1 <dbl [1,000]>
## 5          5     1 <dbl [1,000]>
## 6          6     1 <dbl [1,000]>

Your data should have the following information:

• subject_id identifies a subject's ID within their group. For example, the number 1 will identify the first person in a group, so there should be multiple 1 in this column, and so on.
• group identifies the group ID that a subject is in
• Xt contains the activity data for each subject, nested for each

After getting your data into this format, you will need to use the add_helper_columns() and add_activity_profile() functions to further process the data. add_helper_columns() helps add some useful columns for later analysis, and add_activity_profile() creates a column that converts the raw activity data into the activity profile that is the crux of the test.

processed_data = example_activity %>% 
    add_helper_columns(group_col = group) %>% 
    add_activity_profile(group_col = group, 
                         activity_col = Xt)

You will need to provide the names of the columns that contain this corresponding information in your data (subject ID, group ID, activity data). Once your data is in this format, it can be passed into the EL_test() function to perform the test.

# Test uses bootstrapping, so set seed for consistency
set.seed(1)

# Conduct the functional ANOVA test
el = processed_data %>% 
  EL_test(id_col = subject_id,
          group_col = group,
          activity_col = Xt)

The test uses bootstrap and uses optimization, so it can take a while to run. See vignettes("demo") for more details.