plot.remify: plot.remify
In remify: Processing and Transforming Relational Event History Data

plot.remify

R Documentation

plot.remify

Description

several plots that describe the network of relational events, both for directed and undirected relational events.

Usage

## S3 method for class 'remify'
plot(
  x,
  which = c(1:10),
  breaks = 15L,
  palette = "Purples",
  n_intervals = 10L,
  rev = TRUE,
  actors = NULL,
  pch.degree = 20,
  igraph.edge.color = "#4daa89",
  igraph.vertex.color = "#5AAFC8",
  ...
)

Arguments

`x`	is a `remify` object.
`which`	one or more numbers between 1 and 7. Plots described in order: (1) distribution of the inter-event times (histogram), (2) tile plot titled 'activity plot', with in-degree and out-degree activity line plots on the sides (or total-degree on the top side if the network is undirected). Tiles' color is scaled based on the count of the directed (or undirected) dyad, (3) for directed networks two plots of normalized out-degree and in-degree (values ranging in [0,1]) over a set of `n_intervals` (evenly spaced). For undirected networks one plot of normalized total-degree over the `n_intervals` (also here values ranging in [0,1]). The normalization is calculated in each interval as the `(degree-min(degree))/(max(degree)-min(degree)))` for each actor considering minimum and maximum degree (in-, out- or total-) observed in the interval (opacity and size of the points is proportional to the normalized measure), (4) four plots: (i) number of events (# events) per time interval, (ii) proportion of observed dyads (# dyads / x$D) per time interval, (iii) and (iv) (for directed network only) proportion of active senders and receivers per time interval (calculated as # senders/ x$N and # receiver/x$N per interval), (5) two networks: (i) network of events where edges are considered undirected (edges' opacity is proportional to the counts of the undirected events, vertices' opacity is proportional to the total-degree of the actors), (ii) visualization of directed network (edges' opacity is proportional to the counts of the directed events, vertices' opacity is proportional to the in-degree of the actors), (6) counting process N(t): cumulative number of events against real (non-ordinal) time, with a dashed constant-rate reference line. A roughly straight curve indicates a stationary baseline rate; systematic departures or kinks indicate non-stationarity. Skipped for ordinal event histories, (7) dyad event raster: one row per dyad with a tick per event, capped at the top 30 most active dyads. For directed histories the two directions of a pair are kept as adjacent rows, so reciprocity appears as neighbouring rows firing in sequence and inertia as horizontal clustering of ticks, (8) distribution of dyad recurrence times: pooled gaps between successive events on the same dyad, with a dashed same-mean exponential reference. Excess mass at short gaps (above the reference) indicates temporal clustering, suggesting window/decay memory effects rather than plain inertia. For directed histories recurrence is measured on directed dyads; for ordinal histories the gap is the event-index distance and the exponential reference is omitted, (9) dyad concentration (Lorenz curve): cumulative share of events against cumulative share of dyads, taken over all `x$D` possible dyads so that never-active dyads enter as zeros. The dashed diagonal is perfect equality; a curve bowing toward the bottom-right indicates events concentrated on few dyads, which motivates frailty and makes per-dyad effects harder to identify. A Gini coefficient is reported in the corner, (10) reciprocity scatter (directed histories only): for each unordered pair, the event count in the busier direction against the count in the quieter direction, with point area proportional to the pair's total volume. Points near the dashed diagonal indicate balanced exchange (reciprocity has little asymmetry to explain); points near the horizontal axis indicate one-directional flow (reciprocity is informative), (11) distribution of event durations (`remify_durem` objects only): histogram of spell lengths (`end - time`) with a same-mean exponential reference for interval histories; for ordinal histories the duration is an event-index gap and the reference is omitted, (12) concurrency over time (`remify_durem` only): number of simultaneously-active spells as a step function, with the peak marked; spells are counted as active over the closed interval `[time, end]`, (13) event spells / Gantt (`remify_durem` only): one lane per dyad with a segment per spell from start to end, capped at the 30 most active dyads and coloured by event type when the history is typed. Note, plots 2–4 use multi-panel layouts and need a sufficiently large graphics device. If you see "figure margins too large", enlarge the plot pane or open a larger device with e.g. `dev.new()`.
`breaks`	default is `15L` and it describes the number of cells of the histogram plot for the inter-event times. It can be specified in the same way as the argument used by the function `graphics::hist()` (see ?graphics::hist for more details).
`palette`	a palette from `grDevices::hcl.pals()` (default is the `"Purples"` palette).
`n_intervals`	number of time intervals for time plots (default is `10`).
`rev`	default is TRUE (reverse order of the color specified in `palette`)
`actors`	default is the set of actors in the network (see `.dict[["actors"]]`). The user can specify a subset of actors on which to run the descriptive plots. If the set contains more than 50 actors, then the function will select the 50 most active actors from the set provided.
`pch.degree`	default is 20. Shape of the points for the degree plots (in-degree, out-degree, total-degree).
`igraph.edge.color`	color of the edges in visualization of the network with vertices and nodes. The user can specify the hex value of a color, the color name or use the function`grDevices::rgb()` which returns the hex value.
`igraph.vertex.color`	color of the vertices in visualization of the network with vertices and nodes. The user can specify the hex value of a color, the color name or use the function `grDevices::rgb()` which returns the hex value.
`...`	other graphical parameters