The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
This Vignette will help you making sure that the data you provide to DigiRhythm follows the library’s guideline. We will also make tackle few useful utility functions that might be helpful, especially:
knitr::opts_chunk$set(echo = TRUE)
library(digiRhythm)
# A sample dataset could be found here
url <- "https://raw.githubusercontent.com/nasserdr/digiRhythm_sample_datasets/main/516b_2.csv"
destination <- file.path(tempdir(), "516b_2.csv")
download.file(url, destfile = destination)
# system(paste("head -n 15", filename)) #Run it only on linux
# IceTag ID:,,50001962,,,,
# Site ID:,,n/a,,,,
# Animal ID:,,n/a,,,,
# First Record:,,30.04.2020,11:54:20,,,
# Last Record:,,15.06.2020,11:06:55,,,
# File Time Zone:,,W. Europe Standard Time,,,,
#
# Date,Time,Motion Index,Standing,Lying,Steps,Lying Bouts
# 30.04.2020,11:54:20,0,0:00.0,0:40.0,0,0
# 30.04.2020,11:55:00,0,0:00.0,1:00.0,0,0
# 30.04.2020,11:56:00,0,0:00.0,1:00.0,0,0
# 30.04.2020,11:57:00,0,0:00.0,1:00.0,0,0
# 30.04.2020,11:58:00,0,0:00.0,1:00.0,0,0
# 30.04.2020,11:59:00,0,0:00.0,1:00.0,0,0
# 30.04.2020,12:00:00,0,0:00.0,1:00.0,0,0
As shown in the previous lines, there are some information that are not needed for working with the data, namely, the first 7 lines. The date and time columns are in two separate columns. In this case, the import_raw_activity_data function is useful. We did not really want to make a universal import function because there might be an unlimited number of cases. However, we only create a function called import_raw_activity_data that solves a couple of most encountered problems, namely the following:
For our particular example, as mentionned in the function call below, we would like to skip 7 lines, we need to read 4 columns, we specified the date and time formats, the separator, the original and target time zones, the sampling rate and we want to remove all days that contains less than 80% of data.
data <- import_raw_activity_data(destination,
skipLines = 7,
act.cols.names = c("Date", "Time", "Motion Index", "Steps"),
date_format = "%d.%m.%Y",
time_format = "%H:%M:%S",
sep = ",",
original_tz = "CET",
target_tz = "CET",
sampling = 15,
trim_first_day = TRUE,
trim_middle_days = TRUE,
trim_last_day = TRUE,
verbose = TRUE
)
#> [1] "Reading the CSV file /tmp/RtmpJSkqhe/516b_2.csv"
#> Removing the following columns because they are not numeric
#> [1] "First data points ... "
#> datetime Motion.Index Steps
#> 1 2020-04-30 11:54:20 0 0
#> 2 2020-04-30 11:55:00 0 0
#> 3 2020-04-30 11:56:00 0 0
#> [1] "Last data point ... "
#> datetime Motion.Index Steps
#> 1 2020-06-15 11:06:00 0 0
#> 2 2020-06-15 11:05:00 2 0
#> 3 2020-06-15 11:04:00 0 0
#> [1] "Minimum Required number of samples per day 76"
#> [1] "Returning a data frame with datetime colum and 2 variable colums"
#> [1] "Total number of samples is 4320 - Total number of days is 45"
As Shown, the argument verbose = TRUE outputs some useful information about the data loading process. We particularly mention the following output: Minimum Required number of samples per day 76. The 76 is obtained by computing 80% of the hypothetical data samples. In fact, if the sampling period is 15 min, then we are supposed to have 96 samples per day. 80% of the 96 samples is equal to 76, therefore, days with less than 96 data points are removed.
Imposing a specific format on the dataset is a design choice. We wanted to avoid future possible errors that might happen when user use the library. A digiRhythm friendly dataset has the following properties: - It has at least two columns. - The first column should be of a POSIX format (datetime). - All other columns should be numeric. - The dataset should contains data for 7 consecutive days at least (7 days are required to compute the DFC).
The function [dgm_friendly()] returns a boolean. If verbose is set to TRUE, it will output details about why or why not the data is not considered digiRhythm friendly.
is_dgm_friendly(data, verbose = TRUE)
#> v Correct time format: First column has a POSIXct Format
#> v Number of days good for DFC: 46 days >= 2 days
#> v Correct numeric format - Column 2 ==> Motion.Index
#> v Correct numeric format - Column 3 ==> Steps
#> The data is digiRhythm friendly
#> [1] TRUE
Sometimes, we may have outliers and want to remove them. Although the outlier removal routine is not complicated, but providing an off-the-shelve function for DGM friendly data is useful and might save time. The [remove_activity_outliers()] (available inside utils.R) will remove the outliers in all the non-POSIX columns. All data points that lie out of the 2.5 and 97.5 percentiles are considered outliers.
Often, data are acquired with milliseconds or seconds sampling periods. However, transforming data to higher sampling periods is recommended for some algorithms like the degree of functional coupling. The library offers resampling function as it is useful in this case. However, here are two considerations:
Users can have access to the periodicity of a dataset using the dgm_periodicity function as shown below: - The output of the function shows the periodicity and start- and end-dates of the data.
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.