In epiverse-trace/linelist: Tagging and Validating Epidemiological Data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

linelist: Tagging and Validating Epidemiological Data

linelist provides a safe entry point to the Epiverse software ecosystem, adding a foundational layer through tagging, validation, and safeguarding epidemiological data, to help make data pipelines more straightforward and robust.

Installation

Stable version

Our stable versions are released on CRAN, and can be installed using:

install.packages("linelist", build_vignettes = TRUE)

::: {.pkgdown-devel}

Development version

The development version of linelist can be installed from GitHub with:

if (!require(pak)) {
  install.packages("pak")
}
pak::pak("epiverse-trace/linelist")

:::

Usage

#| fig.alt: "Graphical summary of the linelist R package, with emphasis of these 4 key features: 1. Tag key epi variables, 2. Validate tagged data, 3. Safeguards vs accidental loss / alteration, 4. Robust data for stronger pipelines](man/figures/linelist_infographics.png"
#| out.width: "60%"
knitr::include_graphics("man/figures/linelist_infographics.png")

linelist works by tagging key epidemiological data in a data.frame or a tibble to facilitate and strengthen data pipelines. The resulting object is a linelist object, which extends data.frame (or tibble) by providing three types of features:

1. a tagging system to identify key data, enabling access to these data using their tags rather than actual names, which may change over time and across datasets

2. validation of the tagged variables (making sure they are present and of the right type/class)

3. safeguards against accidental losses of tagged variables in common data handling operations

The short example below illustrates these different features. See the Documentation section for more in-depth examples and details about linelist objects.

# load packages and a dataset for the example
# -------------------------------------------
library(linelist)
library(dplyr)

dataset <- outbreaks::mers_korea_2015$linelist
head(dataset)

# check known tagged variables
# ----------------------------
tags_names()

# build a linelist
# ----------------
x <- dataset %>%
  tibble() %>%
  make_linelist(
    date_onset = "dt_onset", # date of onset
    date_reporting = "dt_report", # date of reporting
    occupation = "age" # mistake
  )
x
tags(x) # check available tags

validate_linelist() will error if one of your tagged column doesn't have the correct type:

# validation of tagged variables
# ------------------------------
## (this flags a likely mistake: occupation should not be an integer)
validate_linelist(x)

# change tags: fix mistakes, add new ones
# ---------------------------------------
x <- x %>%
  set_tags(
    occupation = NULL, # tag removal
    gender = "sex", # new tag
    outcome = "outcome"
  )

# safeguards against actions losing tags
# --------------------------------------
## attemping to remove geographical info but removing dates by mistake
x_no_geo <- x %>%
  select(-(5:8))

For stronger pipelines, you can even trigger errors upon loss:

lost_tags_action("error")

x_no_geo <- x %>%
  select(-(5:8))

x_no_geo <- x %>%
  select(-(5:7))

## to revert to default behaviour (warning upon error)
lost_tags_action()

Alternatively, content can be accessed by tags:

x_no_geo %>%
  select(has_tag(c("date_onset", "outcome")))

x_no_geo %>%
  tags_df()

linelist can also be connected to the incidence2 package for pipelines focused on aggregated count data:

library(incidence2)

x_no_geo %>%
  tags_df() %>%
  incidence("date_onset", groups = c("gender", "outcome")) %>%
  plot(
    fill = "outcome",
    angle = 45,
    nrow = 2,
    border_colour = "white",
    legend = "bottom"
  )

Documentation

More detailed documentation can be found at: https://epiverse-trace.github.io/linelist/

In particular:

• A general introduction to linelist

• The reference manual

Getting help

To ask questions or give us some feedback, please use the github issues system.

Data privacy

Case line lists may contain personally identifiable information (PII). While linelist provides a way to store this data in R, it does not currently provide tools for data anonymization. The user is responsible for respecting individual privacy and ensuring PII is handled with the required level of confidentiality, in compliance with applicable laws and regulations for storing and sharing PII.

Note that PII is rarely needed for common analytics tasks, so that in many instances it may be advisable to remove PII from the data before sharing them with analytics teams.

Development

Lifecycle

This package is currently maturing, as defined by the RECON software lifecycle. This means that essential features and mechanisms are present and stable but minor breaking changes, or function renames may still occur sporadically.

Contributions

Contributions are welcome via pull requests.

Code of Conduct

Please note that the linelist project is released with a Code of Conduct. By contributing to this project, you agree to abide by its terms.

Notes

This package is a reboot of the RECON package linelist. Unlike its predecessor, the new package focuses on the implementation of a linelist class. The data cleaning features of the original package will eventually be re-implemented for linelist objects, albeit likely in a separate package.

epiverse-trace/linelist documentation built on May 11, 2024, 2:23 p.m.