panelView – User Manual

Welcome

This manual serves as a user guide for the panelView package in R, which visualizes panel data. For the technical details, see Mou, Liu, and Xu (2022).

Installation

Source	Version	Date	Features
CRAN	1.3.1	2026-05-13	Refreshed plot defaults; `type = "network"`; `theme = "red"`; `group.mean.overlay`
GitHub (`master`)	1.3.1	2026-05-13	Same as CRAN
GitHub (`dev`)	1.3.2	2026-05-14	+ reversal-gap fix; dotted bridges for missing observations under `leave.gap = FALSE`

# From CRAN (stable release)
install.packages("panelView")

# Stable GitHub version (master branch)
devtools::install_github("xuyiqing/panelView")

# Development version (dev branch, includes type = "network")
devtools::install_github("xuyiqing/panelView@dev")

# Check installed version
installed.packages()["panelView", "Version"]
#> [1] "1.3.3"

panelView depends on the following packages, which should be installed automatically when panelView is being installed. You can also install them manually.

install_all <- function(packages) {
  installed_pkgs <- installed.packages()[, "Package"]
  for (pkg in packages) {
    if (!pkg %in% installed_pkgs) {
      install.packages(pkg)
    }
  }
}
packages <- c("ggplot2", "gridExtra", "grid", "dplyr")
install_all(packages)

The type = "network" plot additionally requires igraph, which is an optional (suggested) dependency:

install.packages("igraph")

Quick-reference

panelView has one main function, panelview(), with four core capabilities:

Treatment status & missingness (type = "treat", type = "missing") — plot which units are treated in which periods, highlight missing values, and summarize treatment histories.
Outcome trajectories (type = "outcome") — plot the raw outcome variable as time series, colored by treatment status.
Bivariate relationships (type = "bivariate") — visualize the relationship between a treatment variable and an outcome variable, either in aggregate or unit by unit.
Network structure (type = "network") — visualize the connectivity structure of panel data as a $k$-partite graph, identifying singletons, connected components, and duplicate observations. Supports two or more sets of fixed effects. See Correia (2016) for the graph-theoretic framework.

Argument	Type	Description
`formula`	formula	`Y ~ D + X1 + X2`; first RHS variable is the treatment
`data`	data.frame	Long-format panel data
`index`	character	`c("unit_var", "time_var")`; length $\geq 2$ for `type = "network"`
`type`	character	`"treat"` / `"missing"` / `"outcome"` / `"bivariate"` / `"network"`
`Y`, `D`	character	Alternative to formula: variable names as strings
`by.timing`	logical	Sort units by timing of first treatment
`pre.post`	logical	Distinguish pre- and post-treatment periods (binary D only)
`collapse.history`	logical	Collapse units with identical treatment histories
`show.id`	integer	Plot only units ranked `show.id` (by alphabetical order)
`id`	vector	Plot only units whose ID matches `id`
`by.group`	logical	Split outcome plot by treatment-status group
`by.cohort`	logical	Collapse outcome by treatment-cohort (type=“outcome” only)
`by.unit`	logical	Plot Y and D per unit (type=“bivariate” only)
`outcome.type`	character	`"continuous"` (default) or `"discrete"`
`leave.gap`	logical	Show time gaps as white bars
`axis.lab.angle`	numeric	Angle of x-axis labels: 0, 45, or 90
`legend.labs`	character	Override legend labels
`color`	character	Override fill/line colors
`legendOff`	logical	Remove the legend
`gridOff`	logical	Remove grid lines (continuous treatment)
`theme`	character	`"default"` (blue) or `"red"` (high-contrast publication recipe)
`group.mean.overlay`	logical	Overlay group-mean line + 10–90% ribbon (`type = "outcome"`)

Note that Y, D, and X are merely labels; they can be any variables in a panel dataset.

Shipped datasets

panelView ships three datasets for examples and testing:

Dataset	Description	Units	Periods	Balance
`turnout`	US state voter turnout and election reform policies	47 states	24 elections (1920–2000)	Balanced
`capacity`	State institutional capacity across countries	47 countries	44 years (1960–2003)	Balanced
`simdata`	Simulated panel with staggered binary treatment	200 units	15 periods	Balanced

All three are balanced panels. The network structure chapter uses inline simulation code to construct unbalanced panels that showcase singletons, connected components, and multi-way fixed effects.

Organization

Chapter 1 Treatment Status — Plotting treatment status and missingness
Chapter 2 Outcome Trajectories — Plotting outcome trajectories
Chapter 3 Bivariate Relationships — Plotting bivariate relationships
Chapter 4 Network Structure — Visualizing network structure: singletons, connected components, and multi-way fixed effects

Contributors

Yiqing Xu
Licheng Liu
Hongyu Mou
StatsClaw (Agentic System for Statistical Software Development)

Report bugs

Please report bugs to yiqingxu [at] stanford.edu or open an issue on GitHub.

# Welcome {.unnumbered} This manual serves as a user guide for the **panelView** package in R, which visualizes panel data. For the technical details, see @mou2022. ## Installation | Source | Version | Date | Features | |--------|---------|------|----------| | CRAN | 1.3.1 | 2026-05-13 | Refreshed plot defaults; `type = "network"`; `theme = "red"`; `group.mean.overlay` | | GitHub (`master`) | 1.3.1 | 2026-05-13 | Same as CRAN | | GitHub (`dev`) | 1.3.2 | 2026-05-14 | + reversal-gap fix; dotted bridges for missing observations under `leave.gap = FALSE` | ```{r install-cran, eval = FALSE} # From CRAN (stable release) install.packages("panelView") ``` ```{r install-github-stable, eval = FALSE, message = FALSE, warning = FALSE, cache = FALSE} # Stable GitHub version (master branch) devtools::install_github("xuyiqing/panelView") ``` ```{r install-github-dev-branch, eval = FALSE, message = FALSE, warning = FALSE, cache = FALSE} # Development version (dev branch, includes type = "network") devtools::install_github("xuyiqing/panelView@dev") ``` ```{r check-version} # Check installed version installed.packages()["panelView", "Version"] ``` **panelView** depends on the following packages, which should be installed automatically when **panelView** is being installed. You can also install them manually. ```{r install-deps, eval = FALSE} install_all <- function(packages) { installed_pkgs <- installed.packages()[, "Package"] for (pkg in packages) { if (!pkg %in% installed_pkgs) { install.packages(pkg) } } } packages <- c("ggplot2", "gridExtra", "grid", "dplyr") install_all(packages) ``` The `type = "network"` plot additionally requires **igraph**, which is an optional (suggested) dependency: ```{r install-igraph, eval = FALSE} install.packages("igraph") ``` ## Quick-reference **panelView** has one main function, `panelview()`, with four core capabilities: 1. **Treatment status & missingness** (`type = "treat"`, `type = "missing"`) — plot which units are treated in which periods, highlight missing values, and summarize treatment histories. 2. **Outcome trajectories** (`type = "outcome"`) — plot the raw outcome variable as time series, colored by treatment status. 3. **Bivariate relationships** (`type = "bivariate"`) — visualize the relationship between a treatment variable and an outcome variable, either in aggregate or unit by unit. 4. **Network structure** (`type = "network"`) — visualize the connectivity structure of panel data as a $k$-partite graph, identifying singletons, connected components, and duplicate observations. Supports two or more sets of fixed effects. See @correia2016 for the graph-theoretic framework. | Argument | Type | Description | |---|---|---| | `formula` | formula | `Y ~ D + X1 + X2`; first RHS variable is the treatment | | `data` | data.frame | Long-format panel data | | `index` | character | `c("unit_var", "time_var")`; length $\geq 2$ for `type = "network"` | | `type` | character | `"treat"` / `"missing"` / `"outcome"` / `"bivariate"` / `"network"` | | `Y`, `D` | character | Alternative to formula: variable names as strings | | `by.timing` | logical | Sort units by timing of first treatment | | `pre.post` | logical | Distinguish pre- and post-treatment periods (binary D only) | | `collapse.history` | logical | Collapse units with identical treatment histories | | `show.id` | integer | Plot only units ranked `show.id` (by alphabetical order) | | `id` | vector | Plot only units whose ID matches `id` | | `by.group` | logical | Split outcome plot by treatment-status group | | `by.cohort` | logical | Collapse outcome by treatment-cohort (type="outcome" only) | | `by.unit` | logical | Plot Y and D per unit (type="bivariate" only) | | `outcome.type` | character | `"continuous"` (default) or `"discrete"` | | `leave.gap` | logical | Show time gaps as white bars | | `axis.lab.angle` | numeric | Angle of x-axis labels: 0, 45, or 90 | | `legend.labs` | character | Override legend labels | | `color` | character | Override fill/line colors | | `legendOff` | logical | Remove the legend | | `gridOff` | logical | Remove grid lines (continuous treatment) | | `theme` | character | `"default"` (blue) or `"red"` (high-contrast publication recipe) | | `group.mean.overlay` | logical | Overlay group-mean line + 10--90% ribbon (`type = "outcome"`) | Note that *Y*, *D*, and *X* are merely labels; they can be any variables in a panel dataset. ## Shipped datasets **panelView** ships three datasets for examples and testing: | Dataset | Description | Units | Periods | Balance | |---------|-------------|-------|---------|---------| | `turnout` | US state voter turnout and election reform policies | 47 states | 24 elections (1920--2000) | Balanced | | `capacity` | State institutional capacity across countries | 47 countries | 44 years (1960--2003) | Balanced | | `simdata` | Simulated panel with staggered binary treatment | 200 units | 15 periods | Balanced | All three are balanced panels. The [network structure chapter](#sec-network) uses inline simulation code to construct unbalanced panels that showcase singletons, connected components, and multi-way fixed effects. ## Organization - [Chapter @sec-treat] — Plotting treatment status and missingness - [Chapter @sec-outcome] — Plotting outcome trajectories - [Chapter @sec-bivariate] — Plotting bivariate relationships - [Chapter @sec-network] — Visualizing network structure: singletons, connected components, and multi-way fixed effects ## Contributors - [Yiqing Xu](https://yiqingxu.org/){target="_blank"} - [Licheng Liu](https://liulch.github.io/){target="_blank"} - [Hongyu Mou](https://hongyumou.github.io/){target="_blank"} - [StatsClaw](https://statsclaw.ai/){target="_blank"} (Agentic System for Statistical Software Development) ## Report bugs Please report bugs to yiqingxu [at] stanford.edu or open an issue on [GitHub](https://github.com/xuyiqing/panelview/issues).  [![Lifecycle: stable](https://lifecycle.r-lib.org/articles/figures/lifecycle-stable.svg)](https://lifecycle.r-lib.org/articles/stages.html#stable) [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT) [<img src="https://www.r-pkg.org/badges/version/panelView" alt="CRAN status"/>](https://CRAN.R-project.org/package=panelView) [<img src="https://cranlogs.r-pkg.org/badges/grand-total/panelView" alt="downloads: CRAN"/>](https://cran.r-project.org/package=panelView)