用R中的下拉线绘制速率

Question

I haven't used stackoverflow in a very long time so please excuse me if this question is poorly laid out.

I have 31 x-values which are the timestamps 1:45PM to 2:15PM inclusive, as well as different rates that occurred at those time stamps.

time    date.1  date.2
1:45    1.0063  1.005
1:46    1.0067  1.00576
1:47    1.0059  1.00559
1:48    1.00559 1.00532
1:49    1.0062  1.00599
1:50    1.0063  1.00622
1:51    1.005   1.00622
1:52    1.00576 1.00612
1:53    1.0066  1.00611
1:54    1.00532 1.00605
1:55    1.00599 1.00559
1:56    1.00622 1.0062
1:57    1.00612 1.00567
1:58    1.00611 1.00578
1:59    1.00605 1.00589
2:00    1.00559 1.00599
2:01    1.0062  1.00611
2:02    1.00567 1.00612
2:03    1.00578 1.00603
2:04    1.00589 1.00599
2:05    1.00599 1.00598
2:06    1.0062  1.00652
2:07    1.0063  1.00642
2:08    1.005   1.00641
2:09    1.00602 1.00635
2:10    1.00603 1.00589
2:11    1.00611 1.0065
2:12    1.00612 1.00597
2:13    1.00603 1.00608
2:14    1.00599 1.00619
2:15    1.00598 1.00629

I essentially want to plot all these values on the same chart and have drop down lines to the x-axis for each point. The reason I want to do this is because I want to visually see where the rates seem to be most of the time at each individual time stamp.

Right now, I am using:

plot(x,y1,type = "h")
par(new=TRUE)
plot(x,y2, type = "h")

This doesn't work if I allow x to be something like:

x <- c("1:45", "1:46", "1:47", "1:48", "1:49", "1:50", "1:51", "1:52", "1:53", "1:54", "1:55", "1:56", "1:57", "1:58", "1:59", "2:00", "2:01", "2:02", "2:03", "2:04", "2:05", "2:06", "2:07", "2:08", "2:09", "2:10", "2:11", "2:12", "2:13", "2:14", "2:15")

I think this isn't working because R isn't recognizing these as values so can't put an xlim on the plot?

I am also going to have several y values (probably 15+) where I plan to store them in excel. (i've just been copying from my clipboard right now)

read.table("clipboard")

I would greatly appreciate anyone's help in showing me how to get the drop-down lines to work and how to get this to be a visually appealing plot!

EDIT:

Thank you for both your answers. I tried both of your suggestions and they both work perfectly. However, I understand why it might be a better idea to just look at the points rather looking at drop down lines, but I am still curious as to what a chart with the drop down lines would look like. I'd like them to be somewhat transparent like was suggested. This is the code i think i might go with simply because I think the for() loop will make it easier for me if there's a lot of variables.

setwd("/Users/null/Desktop")
data <- read.table("Workbook2.csv", header=TRUE, sep=",")

#this will fix the issue of trying to use strings
data$x <- seq_along(data$time)

#Now plot them. Start by making an empty plot space that covers the full range
# of y values in your data set; then make a better x-axis; then plot the points, using
# alpha() from 'scales' to make the points transparent so overlapping ones show up.
with(data, plot(x = x, y = date.1, type = "n", xaxt="n", pch=20, bg=alpha("black", 0.5), col=alpha("black", 0.5),
          ylab = "Rate", xlab = "Time", ylim = c(min(data[,2:3]), max(data[,2:3]))))
axis(1, at=seq(length(data$x)), labels=data$time, tick=FALSE, las=2)
for (i in 1:2) points(x = data$x, y = data[,paste0("date.", i)], pch=20, col=alpha("black", 0.5))

Perhaps I should consider also adding a line across all the data that shows the average rate in the variables for each time stamp?

Answer 1

You're trying to plot numbers against strings, and that's why it doesn't work.

Suppose your data frame is called dz , then:

dz$time<-as.POSIXct(dz$time,format="%H:%M")

will convert your strings to actual times, from where you can plot:

library(ggplot2)
ggplot(dz,aes(time,y1,color="blue"))+geom_point()+geom_point(aes(y=y2,color="red"))

If you want to visually determine the times at which the y1 , y2 variables changed drastically, then the use of geom_line instead of geom_point could be more appropriate.

Answer 2

If you want to see how rates cluster (or not) at each time slice, I think you'd do better to use points rather than vertical lines, ideally with transparent colors so you can see where they overlap. Here's one way to do that in base R, using a toy version of your data set with five y series:

# Make a toy data set with five y series
set.seed(1)
df <- data.frame(time = c(paste(rep(1, 15), seq(45, 59, 1), sep=":"), paste0(rep(2, 10), ":0",  seq(0, 9)), paste(rep(2, 5), seq(10, 15), sep=":")),
    y1 = rnorm(31, 1, 0.1), y2 = rnorm(31, 1, 0.1), y3 = rnorm(31, 1, 0.1), y4 = rnorm(31, 1, 0.1), y5 = rnorm(31, 1, 0.1), stringsAsFactors=FALSE)

# Make a sequence along your time series to use as your x values; as the previous
# answer said, you can't use raw strings as x values for plotting.
df$x <- seq_along(df$time)

# Now plot them. Start by making an empty plot space that covers the full range
# of y values in your data set; then make a better x-axis; then plot the points, using
# alpha() from 'scales' to make the points transparent so overlapping ones show up.
with(df, plot(x = x, y = y1, type = "n", xaxt="n", pch=20, bg=alpha("black", 0.5), col=alpha("black", 0.5),
    ylab = "y value", xlab = "time", ylim = c(min(df[,2:6]), max(df[,2:6]))))
axis(1, at=seq(length(df$x)), labels=df$time, tick=FALSE, las=2)
for (i in 1:5) points(x = df$x, y = df[,paste0("y", i)], pch=20, col=alpha("black", 0.5))

Here's what I get with that. It will be more effective with data that aren't randomly generated, but I think you get the idea:

And if you want vertical lines to make it easier to tie the dots to the time stamps, add something like abline(v = df$x, col = "gray85", lwd=0.5) before you plot the points. Here's what that produces: