--- title: "Explore missing data" output: rmarkdown::html_vignette: toc: true toc_depth: 4 number_sections: true vignette: > %\VignetteIndexEntry{Explore missing data} %\VignetteEncoding{UTF-8} %\VignetteEngine{knitr::rmarkdown} --- ```{r message=FALSE, warning=FALSE, include=FALSE} library(REDCapDM) library(kableExtra) library(knitr) library(dplyr) library(purrr) ``` The examples use `covican`, the dataset included in the package (described in `vignette("REDCapDM")`). # **rd_missplot** The `rd_missplot()` function gives a visual summary of the missing values of the dataset. It needs the `ggplot2` package. By default (`by = "vars"`), it shows the percentage of missing values of each variable. In longitudinal projects, it is a heatmap with one row per event. When the event-form mapping is available, each variable is only evaluated in the events where its form is collected, and the other cells are left blank: ```{r message=FALSE, warning=FALSE, fig.width=8, fig.height=5, eval=requireNamespace("ggplot2", quietly = TRUE)} rd_missplot(covican) ``` With `by = "record"`, the plot shows one row per record and one column per variable, with the missing values painted. In longitudinal projects there is one panel per event, each with the variables collected in that event. By default only the first event is plotted (with a warning), so use the `events` argument to choose others (`events = "all"` for every event). Records without data in an event are shown in grey. In this example we use the first 30 records, so that their identifiers can be read: ```{r message=FALSE, warning=FALSE, fig.width=8, fig.height=7, eval=requireNamespace("ggplot2", quietly = TRUE)} covican_30 <- covican covican_30$data <- covican$data |> dplyr::filter(record_id %in% unique(record_id)[1:30]) rd_missplot(covican_30, by = "record", events = "all", forms = c("vital_signs", "laboratory_findings")) ``` Other useful arguments: - `events` and `forms`: plot only some events or forms, in both modes. - `drop_complete`: hide the variables without missing values (default `TRUE`). - `max_vars`: plot only the variables with the most missing values. - `use_field_labels` and `max_label_length`: use the labels of the dictionary for the variables, shortened to a maximum number of characters (default 40).