Incorporate plotting for Goldfish package

[WylieMabel]

Current:
Plotting methods examine_outliers and examine_changepoints process and plot data in the class_diagnostics.R in the Goldfish package.

Proposed Solution:
Align with other stocnet packages and process only the data in goldfish code, and implement methods in the Autograph package to plot.

[jhollway]

Let me know when these functions export only a class of object (and what the name of the class is) and I'll implement the plots.

[WylieMabel]

Hi James, sorry for the delay, I have separated the object creation from the plotting and pushed these methods to the develop branch in goldfish here.

The functions examine_outliers and examine_changepoints export classes outliers.goldfish and changepoints.goldfish respectively.

I have provided the extracted plotting code for both below, and tests related to the examine methods are found in tests/test-class_diagnostics.R.

`plot_outliers <- function(data) {
if (!"YES" %in% data$outlier) {
cat("No outliers found.\n")
return(invisible(NULL))
}

ggplot2::ggplot(data, ggplot2::aes(x = .data$time, y = .data$intervalLogL)) +
ggplot2::geom_line() +
ggplot2::geom_point(ggplot2::aes(color = .data$outlier)) +
ggplot2::geom_text(ggplot2::aes(label = .data$label),
angle = 270, size = 2,
hjust = "outward", color = "red"
) +
ggplot2::theme_minimal() +
ggplot2::scale_colour_manual(
values = c("black", "red"),
guide = "none"
) +
ggplot2::xlab("") +
ggplot2::ylab("Interval log likelihood")
}

plot_changepoints <- function(result) {
data <- result$data
cpt.pts <- result$cpt_points

if (is.null(cpt.pts) || length(cpt.pts) == 0) {
cat("No regime changes found.\n")
return(invisible(NULL))
}

ggplot2::ggplot(data, ggplot2::aes(x = .data$time, y = .data$intervalLogL)) +
ggplot2::geom_line() +
ggplot2::geom_point() +
ggplot2::geom_vline(
xintercept = na.exclude(data$time[cpt.pts]),
color = "red"
) +
ggplot2::theme_minimal() +
ggplot2::xlab("") +
ggplot2::ylab("Interval log likelihood") +
ggplot2::scale_x_continuous(
breaks = data$time[cpt.pts],
labels = data$time[cpt.pts]
) +
ggplot2::theme(axis.text.x = ggplot2::element_text(angle = 45, hjust = 1))
}`

Let me know if I can help further.

[jhollway]

Thanks @WylieMabel, I've added the plot functions as methods, and will further adapt them to operate as {autograph} functions. In the meantime though, I have one suggestion and one request:

• I would recommend against including a period in the class name, e.g. "outliers.goldfish", because it could cause some dispatching issues later down the road if there is also a "goldfish" class. This can be avoided by using an underscore in place of the period, such that the plot method can be plot.outliers_goldfish().
• Can you please send me a demonstration/test object for each of these new classes that I can include in the {autograph} package and make sure everything is working as it should?

Cheers

[WylieMabel]

No problem @jhollway

Let me know if that is what you were asking for objects,

## dput for a outlier.goldfish object with 5 outliers
structure(list(time = c(1, 6, 9, 13, 15, 16, 19, 23, 28, 29, 
32, 36), sender = c("Actor 1", "Actor 3", "Actor 2", "Actor 2", 
"Actor 5", "Actor 1", "Actor 3", "Actor 3", "Actor 4", "Actor 2", 
"Actor 5", "Actor 1"), receiver = c("Actor 2", "Actor 2", "Actor 3", 
"Actor 3", "Actor 1", "Actor 5", "Actor 4", "Actor 4", "Actor 2", 
"Actor 3", "Actor 2", "Actor 2"), increment = c(1, 2, 1, 1, 1, 
2, 1, 1, 1, 1, 1, 1), intervalLogL = c(-0.881373467194406, -0.000174112099072718, 
-0.000134740334488879, -0.000134736583448288, -1.09861228866811, 
-0.000297181317265376, -0.00022998968572891, -0.00022998968572891, 
-2.34584221468292e-08, -0.693147180559945, -0.881485316929311, 
-0.000325220883833918), label = c("Actor 1-Actor 2", "", "", 
"", "Actor 5-Actor 1", "", "", "", "", "Actor 2-Actor 3", "Actor 5-Actor 2", 
"Actor 1-Actor 2"), outlier = c("YES", "NO", "NO", "NO", "YES", 
"NO", "NO", "NO", "NO", "YES", "YES", "YES")), row.names = c(NA, 
-12L), nodes = "actorsEx", default_network = "networkState", type = "dyadic", class = c("outliers.goldfish", 
"dependent.goldfish", "data.frame"))

## dput for a changepoint.goldfish object with 8 changepoints
structure(list(data = structure(list(time = structure(c(-308880000, 
-300585600, -288057600, -287884800, -278035200, -266803200, -266803200, 
-265680000, -250560000, -247968000, -243388800, -243388800, -226368000, 
-190425600, -180748800, -179539200, -175564800, -165974400, -160963200, 
-160444800, -159321600, -154742400, -145756800, -144028800, -144028800, 
-142905600, -97545600, -95817600, -90979200, -90979200, -87955200, 
-86572800, -85276800, -85276800, -83635200, -83635200, -82771200, 
-79228800, -79142400, -79142400, -78105600, -68860800, -66355200, 
-65923200, -65491200, -64368000, -64368000, -64368000, -63417600, 
-57456000, -54604800, -53136000, -50198400, -34560000, -34041600, 
-33264000, -33177600, -32313600, -32313600, -31276800, -28944000, 
-28944000, -28944000, -27129600, -18662400, -17539200, -17452800, 
-14083200, -1987200), class = c("POSIXct", "POSIXt"), tzone = "GMT"), 
    sender = c("FIN", "FIN", "FIN", "GBR", "GBR", "DEU", "DEU", 
    "DNK", "BGR", "NOR", "NOR", "NOR", "JOR", "GHA", "FRA", "FIN", 
    "FIN", "GBR", "JPN", "DEU", "RUS", "RUS", "FIN", "DEU", "DEU", 
    "JPN", "JPN", "DNK", "RUS", "RUS", "ESP", "NOR", "DNK", "DNK", 
    "JPN", "JPN", "CAN", "BEL", "DNK", "DNK", "JPN", "MEX", "RUS", 
    "DEU", "DNK", "RUS", "RUS", "RUS", "ARG", "JPN", "RUS", "DNK", 
    "DNK", "AUS", "RUS", "BRA", "RUS", "JPN", "JPN", "ESP", "RUS", 
    "RUS", "RUS", "ESP", "IDN", "POL", "FIN", "NOR", "ESP"), 
    receiver = c("SWE", "RUS", "NOR", "NOR", "ISL", "ISL", "ISL", 
    "ISL", "SRB", "RUS", "RUS", "RUS", "SAU", "RUS", "GBR", "RUS", 
    "NOR", "NOR", "USA", "NLD", "USA", "USA", "RUS", "DNK", "DNK", 
    "KOR", "USA", "NOR", "USA", "USA", "FRA", "SWE", "NOR", "NOR", 
    "USA", "USA", "USA", "DNK", "EUU", "EUU", "NZL", "USA", "USA", 
    "DNK", "SWE", "USA", "USA", "USA", "BRA", "MEX", "USA", "NOR", 
    "NLD", "JPN", "USA", "URY", "USA", "USA", "USA", "MAR", "USA", 
    "USA", "USA", "PRT", "PHL", "USA", "RUS", "SWE", "PRT"), 
    increment = c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 
    1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1), intervalLogL = c(-6.58808035768826, 
    -3.90796779289374, -2.9492447879255, -3.08718469064986, -9.06110377202691, 
    -8.40423389374255, -8.40423389374255, -7.30577062204028, 
    -9.601854757565, -6.53760113461079, -7.09147782620283, -7.09147782620283, 
    -7.32335640035485, -9.52519021133067, -3.89487642740493, 
    -4.71876345119945, -3.68012273143274, -3.62216918595031, 
    -6.57053012527784, -9.00218857355948, -2.75801001751782, 
    -1.80893440454637, -4.04370536511843, -5.3552556648367, -5.3552556648367, 
    -7.22966062492067, -3.81126378915605, -3.17638941604777, 
    -1.71974413429232, -1.71974413429232, -6.21686655072442, 
    -4.15342034974066, -3.02717737830867, -3.02717737830867, 
    -3.42403429401563, -3.42403429401563, -3.4530410556602, -6.55160458394115, 
    -10.0839951954468, -10.0839951954468, -10.2690303892233, 
    -4.24748954378411, -1.68484384213778, -4.67444290538394, 
    -5.83018679507679, -1.69336229170686, -1.69336229170686, 
    -1.69336229170686, -8.1638456235971, -8.81947010270636, -1.61191607918256, 
    -1.78704899215799, -6.04817160622155, -8.63454091086823, 
    -1.60846534767802, -8.23158007012582, -1.48900299802998, 
    -2.08559293253744, -2.08559293253744, -8.39530963483573, 
    -1.68428768524767, -1.68428768524767, -1.68428768524767, 
    -7.85014743465059, -8.85081408376917, -9.77337871201207, 
    -6.26983575079562, -4.60921443993169, -6.75374983983816)), row.names = c("36", 
"37", "38", "39", "40", "41", "42", "43", "44", "45", "46", "47", 
"48", "49", "50", "51", "52", "53", "54", "55", "56", "57", "58", 
"59", "60", "61", "62", "63", "64", "65", "66", "67", "68", "69", 
"70", "71", "72", "73", "74", "75", "76", "77", "78", "79", "80", 
"81", "82", "83", "84", "85", "86", "87", "88", "89", "90", "91", 
"92", "93", "94", "95", "96", "97", "98", "99", "100", "101", 
"102", "103", "104"), nodes = "states", default_network = "bilatnet", type = "dyadic", class = c("dependent.goldfish", 
"data.frame")), cpt_points = c(4L, 14L, 38L, 41L, 48L, 56L, 63L, 
69L)), class = c("changepoints.goldfish", "list"))

If what you were requesting is the code in goldfish to build these outputs I can provide.

As for the renaming/not including period, the classes are named that way to follow naming convention within the goldfish package. Perhaps @auzaheta can provide more clarity on alternate names.

[jhollway]

@auzaheta, should 'changepoints' maybe just get rebranded as a time heterogeneity test? And outliers as concerning leverage? Or how should we "generalise" these diagnostic tools/tests for similar ones in other packages.