Related
I have a dataset that I am modeling with a gam. Because there are two continuous varaibles in the gam, I have centred and scaled these variables before adding them to the model. Therefore, when I use the built-in features in gratia to show the results, the x values are not the same as the original scale. I'd like to plot the results using the scale of the original data.
An example:
library(tidyverse)
library(mgcv)
library(gratia)
set.seed(42)
df <- data.frame(
doy = sample.int(90, 300, replace = TRUE),
year = sample(c(1980:2020), size = 300, replace = TRUE),
site = c(rep("A", 150), rep("B", 80), rep("C", 70)),
sex = sample(c("F", "M"), size = 300, replace = TRUE),
mass = rnorm(300, mean = 500, sd = 50)) %>%
mutate(doy.s = scale(doy, center = TRUE, scale = TRUE),
year.s = scale(year, center = TRUE, scale = TRUE),
across(c(sex, site), as.factor))
m1 <- gam(mass ~
s(year.s, site, bs = "fs", by = sex, k = 5) +
s(doy.s, site, bs = "fs", by = sex, k = 5) +
s(sex, bs = "re"),
data = df, method = "REML", family = gaussian)
draw(m1)
How do I re-plot the last two panels in this figure to show the relationship between year and mass with ggplot?
You can't do this with gratia::draw automatically (unless I'm mistaken).* But you can use gratia::smooth_estimates to get a dataframe which you can then do whatever you like with.
To answer your specific question: to re-plot the last two panels of the plot you provided, but with year unscaled, you can do the following
# Get a tibble of smooth estimates from the model
sm <- gratia::smooth_estimates(m1)
# Add a new column for the unscaled year
sm <- sm %>% mutate(year = mean(df$year) + (year.s * sd(df$year)))
# Plot the smooth s(year.s,site) for sex=F with year unscaled
pF <- sm %>% filter(smooth == "s(year.s,site):sexF" ) %>%
ggplot(aes(x = year, y = est, color=site)) +
geom_line() +
theme(legend.position = "none") +
labs(y = "Partial effect", title = "s(year.s,site)", subtitle = "By: sex; F")
# Plot the smooth s(year.s,site) for sex=M with year unscaled
pM <- sm %>% filter(smooth == "s(year.s,site):sexM" ) %>%
ggplot(aes(x = year, y = est, color=site)) +
geom_line() +
theme(legend.position = "none") +
labs(y = "Partial effect", title = "s(year.s,site)", subtitle = "By: sex; M")
library(patchwork) # use `patchwork` just for easy side-by-side plots
pF + pM
to get:
EDIT: If you also want to shift result on the y-axis as #GavinSimpson (who is the author and maintainer of gratia) mentioned, you can do this with add_constant, adding this code before plotting above:
sm <- sm %>%
add_constant(coef(m1)["(Intercept)"]) %>%
transform_fun(inv_link(m1))
[You should also in general untransform the smooth by the inverse of the model's link function. In your case this is just the identity, so it is not necessary, but in general it would be. That's what the second step above is doing.]
In your example, this results in:
*As mentioned in the custom-plotting vignette for gratia, the goal of draw not to be fully customizable, but just to be useful default. See there for recommendations about custom plots.
I have created this plot with 18 grids using facet_grid command and two different fitting equations (for Jan - Apr, and May - Jun). I have two things that I need help with:
(may sound obvious, but) I haven't been able to find on the internet working codes extract a curve maximum for a stat_smooth fit. I'd appreciate if someone could show and explain what the codes mean. This is the closest I could find, but I am not sure what it means:
gb <- ggplot_build(p1)
curve_max <- gb$data[[1]]$x[which(diff(sign(diff(gb$data[[1]]$y)))==-2)+1]
How to add a vertical line to indicate max value on each curve?
Data file (rlc2 <- read_excel)
Plot
plot <- ggplot(rlc2, aes(par, etr, color=month, group=site))+
geom_point()+
stat_smooth(data = subset(rlc2, rlc2$month!="May" & rlc2$month!="Jun"),
method = "glm",
formula = y ~ x + log(x),
se = FALSE,
method.args = list(family = gaussian(link = "log"), start=c(a=0, b=0, c=0)))+
stat_smooth(data = subset(rlc2, rlc2$month=="May" | rlc2$month=="Jun"),
method = "nlsLM",
formula = y ~ M*(1 - exp(-(a*x))),
se = FALSE,
method.args = list(start=c(M=0, a=10)))+
facet_grid(rows = vars(month), cols = vars(site))
plot
field_rlc_plot
Any other advice are also welcome. I am educated as programmer so my codes are probably a bit messy. Thank you for helping.
Try this:
First, fit the data and extract the maximum of the fit.
my.fit <- function(month, site, data) {
fit <- glm(formula = etr ~ par + log(par),
data = data,
family=gaussian(link = "log")
)
#arrange the dersired output in a tibble
tibble(max = max(fit$fitted.values),
site = site,
month = month)
}
#Apply a custom function `my.fit` on each subset of data
#according to month and site using the group_by/nest/map method
# the results are rowbinded and returned in a data.frame
my.max<-
rlc2 %>%
dplyr::group_by(month, site) %>%
tidyr::nest() %>%
purrr::pmap_dfr(my.fit)
Next, join the results back on your data and plot a geom_line
rlc2 %>%
dplyr::left_join(my.max) %>%
ggplot(aes(x = par, y = etr))+
geom_point()+
stat_smooth(data = subset(rlc2, rlc2$month!="May" & rlc2$month!="Jun"),
method = "glm",
formula = y ~ x + log(x),
se = FALSE,
method.args = list(family = gaussian(link = "log"), start=c(a=0, b=0, c=0)))+
stat_smooth(data = subset(rlc2, rlc2$month=="May" | rlc2$month=="Jun"),
method = "nlsLM",
formula = y ~ M*(1 - exp(-(a*x))),
se = FALSE,
method.args = list(start=c(M=0, a=10)))+
geom_line(aes(y=max), col="red")+
facet_grid(rows = vars(month), cols = vars(site))
I'm trying to create a depth profile graph with the variables depth, distance and temperature. The data collected is from 9 different points with known distances between them (distance 5m apart, 9 stations, 9 different sets of data). The temperature readings are according to these 9 stations where a sonde was dropped directly down, taking readings of temperature every 2 seconds. Max depth at each of the 9 stations were taken from the boat also.
So the data I have is:
Depth at each of the 9 stations (y axis)
Temperature readings at each of the 9 stations, at around .2m intervals vertical until the bottom was reached (fill area)
distance between the stations, (x axis)
Is it possible to create a depth profile similar to this? (obviously without the greater resolution in this graph)
I've already tried messing around with ggplot2 and raster but I just can't seem to figure out how to do this.
One of the problems I've come across is how to make ggplot2 distinguish between say 5m depth temperature reading at station 1 and 5m temperature reading at station 5 since they have the same depth value.
Even if you can guide me towards another program that would allow me to create a graph like this, that would be great
[ REVISION ]
(Please comment me if you know more suitable interpolation methods, especially not needing to cut under bottoms data.)
ggplot() needs long data form.
library(ggplot2)
# example data
max.depths <- c(1.1, 4, 4.7, 7.7, 8.2, 7.8, 10.7, 12.1, 14.3)
depth.list <- sapply(max.depths, function(x) seq(0, x, 0.2))
temp.list <- list()
set.seed(1); for(i in 1:9) temp.list[[i]] <- sapply(depth.list[[i]], function(x) rnorm(1, 20 - x*0.5, 0.2))
set.seed(1); dist <- c(0, sapply(seq(5, 40, 5), function(x) rnorm(1, x, 1)))
dist.list <- sapply(1:9, function(x) rep(dist[x], length(depth.list[[x]])))
main.df <- data.frame(dist = unlist(dist.list), depth = unlist(depth.list) * -1, temp = unlist(temp.list))
# a raw graph
ggplot(main.df, aes(x = dist, y = depth, z = temp)) +
geom_point(aes(colour = temp), size = 1) +
scale_colour_gradientn(colours = topo.colors(10))
# a relatively raw graph (don't run with this example data)
ggplot(main.df, aes(x = dist, y = depth, z = temp)) +
geom_raster(aes(fill = temp)) + # geom_contour() +
scale_fill_gradientn(colours = topo.colors(10))
If you want a graph such like you showed, you have to do interpolation. Some packages give you spatial interpolation methods. In this example, I used akima package but you should think seriously that which interpolation methods to use.
I used nx = 300 and ny = 300 in below code but I think it would be better to decide those values carefully. Large nx and ny gives a high resolution graph, but don't foreget real nx and ny (in this example, real nx is only 9 and ny is 101).
library(akima); library(dplyr)
interp.data <- interp(main.df$dist, main.df$depth, main.df$temp, nx = 300, ny = 300)
interp.df <- interp.data %>% interp2xyz() %>% as.data.frame()
names(interp.df) <- c("dist", "depth", "temp")
# draw interp.df
ggplot(interp.df, aes(x = dist, y = depth, z = temp)) +
geom_raster(aes(fill = temp)) + # geom_contour() +
scale_fill_gradientn(colours = topo.colors(10))
# to think appropriateness of interpolation (raw and interpolation data)
ggplot(interp.df, aes(x = dist, y = depth, z = temp)) +
geom_raster(aes(fill = temp), alpha = 0.3) + # interpolation
scale_fill_gradientn(colours = topo.colors(10)) +
geom_point(data = main.df, aes(colour = temp), size = 1) + # raw
scale_colour_gradientn(colours = topo.colors(10))
Bottoms don't match !!I found ?interp says "interpolation only within convex hull!", oops... I'm worrid about the interpolation around the problem-area, is it OK ? If no problem, you need only cut the data under the bottoms. If not, ... I can't answer immediately (below is an example code to cut).
bottoms <- max.depths * -1
# calculate bottom values using linear interpolation
approx.bottoms <- approx(dist, bottoms, n = 300) # n must be the same value as interp()'s nx
# change temp values under bottom into NA
library(dplyr)
interp.cut.df <- interp.df %>% cbind(bottoms = approx.bottoms$y) %>%
mutate(temp = ifelse(depth >= bottoms, temp, NA)) %>% select(-bottoms)
ggplot(interp.cut.df, aes(x = dist, y = depth, z = temp)) +
geom_raster(aes(fill = temp)) +
scale_fill_gradientn(colours = topo.colors(10)) +
geom_point(data = main.df, size = 1)
If you want to use stat_contour
It is harder to use stat_contour than geom_raster because it needs a regular grid form. As far as I see your graph, your data (depth and distance) don't form a regular grid, it means it is much difficult to use stat_contour with your raw data. So I used interp.cut.df to draw a contour plot. And stat_contour have a endemic problem (see How to fill in the contour fully using stat_contour), so you need to expand your data.
library(dplyr)
# 1st: change NA into a temp's out range value (I used 0)
interp.contour.df <- interp.cut.df
interp.contour.df[is.na(interp.contour.df)] <- 0
# 2nd: expand the df (It's a little complex, so please use this function)
contour.support.func <- function(df) {
colname <- names(df)
names(df) <- c("x", "y", "z")
Range <- as.data.frame(sapply(df, range))
Dim <- as.data.frame(t(sapply(df, function(x) length(unique(x)))))
arb_z = Range$z[1] - diff(Range$z)/20
df2 <- rbind(df,
expand.grid(x = c(Range$x[1] - diff(Range$x)/20, Range$x[2] + diff(Range$x)/20),
y = seq(Range$y[1], Range$y[2], length = Dim$y), z = arb_z),
expand.grid(x = seq(Range$x[1], Range$x[2], length = Dim$x),
y = c(Range$y[1] - diff(Range$y)/20, Range$y[2] + diff(Range$y)/20), z = arb_z))
names(df2) <- colname
return(df2)
}
interp.contour.df2 <- contour.support.func(interp.contour.df)
# 3rd: check the temp range (these values are used to define contour's border (breaks))
range(interp.cut.df$temp, na.rm=T) # 12.51622 20.18904
# 4th: draw ... the bottom border is dirty !!
ggplot(interp.contour.df2, aes(x = dist, y = depth, z = temp)) +
stat_contour(geom="polygon", breaks = seq(12.51622, 20.18904, length = 11), aes(fill = ..level..)) +
coord_cartesian(xlim = range(dist), ylim = range(bottoms), expand = F) + # cut expanded area
scale_fill_gradientn(colours = topo.colors(10)) # breaks's length is 11, so 10 colors are needed
# [Note]
# You can define the contour's border values (breaks) and colors.
contour.breaks <- c(12.5, 13.5, 14.5, 15.5, 16.5, 17.5, 18.5, 19.5, 20.5)
# = seq(12.5, 20.5, 1) or seq(12.5, 20.5, length = 9)
contour.colors <- c("darkblue", "cyan3", "cyan1", "green3", "green", "yellow2","pink", "darkred")
# breaks's length is 9, so 8 colors are needed.
# 5th: vanish the bottom border by bottom line
approx.df <- data.frame(dist = approx.bottoms$x, depth = approx.bottoms$y, temp = 0) # 0 is dummy value
ggplot(interp.contour.df2, aes(x = dist, y = depth, z = temp)) +
stat_contour(geom="polygon", breaks = contour.breaks, aes(fill = ..level..)) +
coord_cartesian(xlim=range(dist), ylim=range(bottoms), expand = F) +
scale_fill_gradientn(colours = contour.colors) +
geom_line(data = approx.df, lwd=1.5, color="gray50")
bonus: legend technic
library(dplyr)
interp.contour.df3 <- interp.contour.df2 %>% mutate(temp2 = cut(temp, breaks = contour.breaks))
interp.contour.df3$temp2 <- factor(interp.contour.df3$temp2, levels = rev(levels(interp.contour.df3$temp2)))
ggplot(interp.contour.df3, aes(x = dist, y = depth, z = temp)) +
stat_contour(geom="polygon", breaks = contour.breaks, aes(fill = ..level..)) +
coord_cartesian(xlim=range(dist), ylim=range(bottoms), expand = F) +
scale_fill_gradientn(colours = contour.colors, guide = F) + # add guide = F
geom_line(data = approx.df, lwd=1.5, color="gray50") +
geom_point(aes(colour = temp2), pch = 15, alpha = 0) + # add
guides(colour = guide_legend(override.aes = list(colour = rev(contour.colors), alpha = 1, cex = 5))) + # add
labs(colour = "temp") # add
You want to treat this as a 3-D surface with temperature as the z dimension. The given plot is a contour plot and it looks like ggplot2 can do that with stat_contour.
I'm not sure how the contour lines are computed (often it's linear interpolation along a Delaunay triangulation). If you want more control over how to interpolate between your x/y grid points, you can calculate a surface model first and feed those z coordinates into ggplot2.
I need to adjust a histogram output in shiny with a range values:
(where m is an arbitrary matrix)
#ui.R
sliderInput(inputId="adjust", label="Choose adjacency threshold", value=c(0.001, 0.9), min=0.0001, max=1),
plotOutput("hist")
#server.R
df<-reactive({
idx = m > min(input$adjust) & m < max(input$adjust)
data.frame(
id = row(m)[idx],
value = m[idx])
})
output$hist<-renderPlot({hist(df()$values)})
However this doesn't seem to affect the histogram- it re-renders when I toggle the slider but its the same each time.... it takes a very long time and it just seems to take all of the values in into account?
Does anyone know how to make this work?
when I try to print the slider's min/max value- nothing comes to the page:
#ui.R
verbatimTextOutput("x")
#server
output$x<-renderPrint({min(input$adjust)})
Hence I might be approaching this the completely wrong way... does anyone know how to do this?
FULL EXAMPLE
library(shiny)
runApp(list(ui = fluidPage(sliderInput(inputId="adjust", label="Choose adjacency threshold", value=c(0.001, 0.9), min=0.0001, max=1),
plotOutput("hist")
server=function(input, output){
adjacentmat<-reactive({adjacency(dat)})
data<-reactive({
adj_mat<-adjacentmat()
adj_mat[adj_mat < input$adjust] <- 0
m<-adj_mat
idx = m > min(input$adjust) & m < max(input$adjust)
data.frame(
source = row(m)[idx],
target = col(m)[idx],
corr = m[idx])
})
output$hist<-renderPlot({hist(data()$corr)})
}
)
generate the dat variable with the following code:
library('dplyr')
set.seed(1)
# generate a couple clusters
nodes_per_cluster <- 30
n <- 10
nvals <- nodes_per_cluster * n
# cluster 1 (increasing)
cluster1 <- matrix(rep((1:n)/4, nodes_per_cluster) +
rnorm(nvals, sd=1),
nrow=nodes_per_cluster, byrow=TRUE)
# cluster 2 (decreasing)
cluster2 <- matrix(rep((n:1)/4, nodes_per_cluster) +
rnorm(nvals, sd=1),
nrow=nodes_per_cluster, byrow=TRUE)
# noise cluster
noise <- matrix(sample(1:2, nvals, replace=TRUE) +
rnorm(nvals, sd=1.5),
nrow=nodes_per_cluster, byrow=TRUE)
dat <- rbind(cluster1, cluster2, noise)
colnames(dat) <- paste0('n', 1:n)
rownames(dat) <- c(paste0('cluster1_', 1:nodes_per_cluster),
paste0('cluster2_', 1:nodes_per_cluster),
paste0('noise_', 1:nodes_per_cluster))
This works for me:
library(shiny)
runApp(list(ui = fluidPage(
mainPanel(sliderInput("test", "Select values", value= c(.001,.9), min= 0.0001, max= 1)),
verbatimTextOutput("test2")
),
server = function(input, output, session) {
output$test2 <- renderPrint(min(input$test))
}))
I'm guessing your problem is somewhere in the code you haven't shown us. Can you give the code for the entire running example of your problem?
I use the following function to convert the topicmodels output to JSON output to use in ldavis.
topicmodels_json_ldavis <- function(fitted, corpus, doc_term){
## Required packages
library(topicmodels)
library(dplyr)
library(stringi)
library(tm)
library(LDAvis)
## Find required quantities
phi <- posterior(fitted)$terms %>% as.matrix
theta <- posterior(fitted)$topics %>% as.matrix
vocab <- colnames(phi)
doc_length <- vector()
for (i in 1:length(corpus)) {
temp <- paste(corpus[[i]]$content, collapse = ' ')
doc_length <- c(doc_length, stri_count(temp, regex = '\\S+'))
}
temp_frequency <- inspect(doc_term)
freq_matrix <- data.frame(ST = colnames(temp_frequency),
Freq = colSums(temp_frequency))
rm(temp_frequency)
## Convert to json
json_lda <- LDAvis::createJSON(phi = phi, theta = theta,
vocab = vocab,
doc.length = doc_length,
term.frequency = freq_matrix$Freq)
return(json_lda)
}
but I receive the following error
Error in LDAvis::createJSON(phi = phi, theta = theta, vocab = vocab, doc.length = doc_length, : Length of doc.length not equal
to the number of rows in theta; both should be equal to the number of
documents in the data.
Here is my complete code:
data <- read.csv("textmining.csv")
corpus <- Corpus(DataframeSource(data.frame(data$reasonforleaving)))
# Remove punctuations and numbers because they are generally uninformative.
corpus <- tm_map(corpus, removePunctuation)
corpus <- tm_map(corpus, removeNumbers)
# Convert all words to lowercase.
corpus <- tm_map(corpus, content_transformer(tolower))
# Remove stopwords such as "a", "the", etc.
corpus <- tm_map(corpus, removeWords, stopwords("english"))
# Use the SnowballC package to do stemming.
library(SnowballC)
corpus <- tm_map(corpus, stemDocument)
# remove extra words
toSpace <- content_transformer(function (x , pattern ) gsub(pattern, " ", x))
corpus <- tm_map(corpus, toSpace, "still")
corpus <- tm_map(corpus, toSpace, "also")
# Remove excess white spaces between words.
corpus <- tm_map(corpus, stripWhitespace)
# Inspect the first document to see what it looks like.
corpus[[1]]$content
dtm <- DocumentTermMatrix(corpus)
# remove empty documents
library(slam)
dtm = dtm[row_sums(dtm)>0,]
# Use topicmodels package to conduct LDA analysis.
burnin <- 500
iter <- 1000
keep <- 30
k <- 5
result55 <- LDA(dtm, 5)
ldaoutput = topicmodels_json_ldavis(result55,corpus, dtm)
Do you know why I receive the error?
Thanks
I had the same issue with same code, and found this function here :
topicmodels2LDAvis <- function(x, ...){
post <- topicmodels::posterior(x)
if (ncol(post[["topics"]]) < 3) stop("The model must contain > 2 topics")
mat <- x#wordassignments
LDAvis::createJSON(
phi = post[["terms"]],
theta = post[["topics"]],
vocab = colnames(post[["terms"]]),
doc.length = slam::row_sums(mat, na.rm = TRUE),
term.frequency = slam::col_sums(mat, na.rm = TRUE)
)
}
Much much simpler to use, just put as argument your LDA result :
result55 <- LDA(dtm, 5)
serVis(topicmodels2LDAvis(result55))
Problem
Your problem is in for (i in 1:length(corpus)) in
doc_length <- vector()
for (i in 1:length(corpus)) {
temp <- paste(corpus[[i]]$content, collapse = ' ')
doc_length <- c(doc_length, stri_count(temp, regex = '\\S+'))
}
Remember, you have removed some "empty" documents from your DocumentTermMatrix in dtm = dtm[row_sums(dtm)>0,],
so your vector length here is going to be too big.
Suggestion
You may want to keep a vector of the empty docs around as it will help you not only to generate the JSON but also to go back and forth between your empty and full document sets.
doc.length = colSums( as.matrix(tdm) > 0 )[!empty.docs]
My suggestion assumes you have the full tdm with empty docs in place