Skip to contents

sardine

Overview

Structured Architecture for Research Data Integration and Evaluation

Unlike the acronym suggests, sardine is designed to help access research data without excessive ceremony. It currently has a focus on REDCap and device data, but the architecture is designed to be extensible to other research data platforms in the future.

Key Features

  • REDCap Integration
    • Project Management: Easily connect to REDCap projects using API tokens, with automatic data caching for efficient access.
    • Quality Assessment: Generate comprehensive data quality reports, including missing data analysis and data type validation against REDCap metadata.
    • Table One Generation: Create detailed Table One summaries with options for filtering, stratification, and statistical testing.
    • Retention Tracking: Tools for tracking participant retention and attrition over time, with customizable plotting functions.
    • Safe Data Importing: Import data back into REDCap with built-in safety checks to prevent accidental overwrites, including change analysis and confirmation prompts.
    • Cross-Project Data Piping: Transfer data between multiple REDCap projects with ease, maintaining data integrity through safety features.
  • Device Data Manipulation
    • Withings data handling:
      • Summarise and format Withings data
      • Generate Sleep Regularity Index (SRI) from sleep/wake data.
      • Reformat for use with other packages (e.g., GGIR, sleepreg).

Installation

You can install the development version of sardine from GitHub using devtools. Make sure you have the devtools package installed first.

# install.packages("devtools")
devtools::install_github("jackmanners/sardine")

Quick Start

The full documentation and reference can be found at https://jackmanners.github.io/sardine/. Below are some basic examples to get you started.

1. Create REDCap Project

Using Environment variables for your API token is optional but recommended if working with a single REDCap project regularly. This also helps avoid hardcoding the API credentials in your scripts (if using GitHub or other version control). Specifically, you’ll need REDCAP_URL and REDCAP_TOKEN set as environment variables.

# Create project (tests connection & caches data automatically)
project <- redcap_project()

# Alternatively, specify URL and token directly
project <- redcap_project(
  url = "https://redcap.your-institution.edu/api/",
  token = "YOUR_API_TOKEN"
)

# View project information
project$info()

# Access project data
all_data <- project$data

2. Viewing Project Structure and Attrition Report

By default, reports and data tables are returned as data.frame objects. The DT package is recommended for exploring and searching data more easily using interactive tables.

library(DT)

# View project structure summary (shiny GUI)
project$explore_project_shiny()

# View project metadata
metadata <- project$metadata
datatable(project$metadata, filter = "top")

# Generate and view data quality report
dq_report <- project$generate_data_quality_report()
datatable(dq_report, filter = "top")
# Plot event completion summary
library(ggplot2)
ggplot(get_event_completion_summary(project), aes(x = event_name, fill = completed)) +
  geom_bar(position = "dodge")

3. Withings Sleep Data 

# Load example Withings data
summary_data <- read.csv("withings_summaries.csv")
epoch_data  <- read.csv("withings_epochs.csv")

# Create Withings sleep object
wsleeps <- withings_sleep(summary_data, epoch_data)

# Generate sleep report
report <- wsleeps$report(output = "html", file = "withings_sleep_report")

# Calculate Sleep Regularity Index (SRI) and save raster plot
wsleep_sri <- wsleeps$sri(plot_path = "sri_plot.png")

# Print SRI results and plot
print(wsleep_sri$sri)
print(wsleep_sri$plot)

Contributing

Please read CONTRIBUTING.md for details on our code of conduct and the process for submitting pull requests.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

  • REDCap (Research Electronic Data Capture) is developed and supported by Vanderbilt University
  • Thanks to the R community for the excellent HTTP and data manipulation packages
  • Thanks to Dan Windred for assistance with Sleep Regularity Index (SRI) calculations