library(tidyverse)
library(lubridate)
library(stringr)
df <-
tibble(Date = as.Date(0:364, origin = "2017-07-01"), Value = rnorm(365)) %>%
mutate(Year = str_sub(Date, 1, 4),
MoFloor = floor_date(Date, unit = "month")) %>%
group_by(Year, MoFloor) %>%
mutate(MoAvgValue = mean(Value)) %>%
ungroup() %>%
group_by(Year) %>%
mutate(MinMoFloor = min(MoFloor),
MaxMoFloor = max(MoFloor),
YearAvgValue = mean(MoAvgValue))
#> # A tibble: 365 x 8
#> # Groups: Year [2]
#> Date Value Year MoFloor
#> <date> <dbl> <chr> <date>
#> 1 2017-07-01 -1.83 2017 2017-07-01
#> 2 2017-07-02 -2.13 2017 2017-07-01
#> 3 2017-07-03 1.49 2017 2017-07-01
#> 4 2017-07-04 0.0753 2017 2017-07-01
#> 5 2017-07-05 -0.437 2017 2017-07-01
#> 6 2017-07-06 -0.327 2017 2017-07-01
#> 7 2017-07-07 -1.28 2017 2017-07-01
#> 8 2017-07-08 0.280 2017 2017-07-01
#> 9 2017-07-09 1.24 2017 2017-07-01
#> 10 2017-07-10 0.0921 2017 2017-07-01
#> # ... with 355 more rows, and 4 more
#> # variables: MoAvgValue <dbl>,
#> # MinMoFloor <date>,
#> # MaxMoFloor <date>,
#> # YearAvgValue <dbl>
Let's first plot the data frame above.
ggplot(df, aes(MoFloor, MoAvgValue, group = Year)) +
facet_grid(~Year, scale = "free_x", space = "free_x") +
geom_point()
In my call to the facet_grid() function I added the arguments scale = "free_x" and space = "free_x" to get rid of empty white space on the plots.
When I go ahead and add geom_segment()s based on group_by()d data, the scale = "free_x" and space = "free_x" arguments are negated. The empty white space reappears!
ggplot(df, aes(MoFloor, MoAvgValue, group = Year)) +
facet_grid(~Year, scale = "free_x", space = "free_x") +
geom_point() +
geom_segment(data = df,
aes(x = min(MinMoFloor),
y = YearAvgValue,
xend = max(MaxMoFloor),
yend = YearAvgValue))
My df data frame is grouped by Year. Why doesn't the geom_segment() function recognize this when I enter (for example) the x = min(MinMoFloor) argument? geom_segment() is pulling the min(MinMoFloor) from the global column, instead of the grouped column. How do I geom_segment() to evaluate the MinMoFloor column as grouped data?
Related
I have a chart built with ggridges, as below, for which I would like to reverse the date order:
To do this I've added the c_trans() function, as defined here, but it requires the day in posixct format, which appears to convert the date to a continuous scale, even when I define group as a factor:
ggplot( lengthCounts2 %>% filter(rwi == rwiFilter),
aes( x = len,
fill = date,
group= factor(date)
)
) +
stat_density( aes( weight = normalised,
y = date, # time_trans works with objects of class POSIXct only
height = after_stat(density)
),
geom = 'density_ridges',
position = 'identity',
adjust = 0.1
) +
scale_y_continuous(trans = rev_date) +
theme_minimal() +
theme( plot.title = element_text(hjust=0.5)
) +
# scale_fill_brewer(palette = "Pastel1") +
labs( title = glue("Sampled/normalised packet size distribution for rwi={rwiFilter} ({rwiText})\n"),
x ="Length (bytes)",
y = "Date"
)
So that I end up with which is clearly not what I want. Is there any way to reverse the y axis but keep it discrete?
Input data:
> lengthCounts2
# A tibble: 8,724 x 5
# Groups: date, rwi [6]
date rwi len n normalised
<dttm> <chr> <dbl> <int> <dbl>
1 2022-04-13 00:00:00 01 35 677 0.0000319
2 2022-04-13 00:00:00 01 40 3113138 0.147
3 2022-04-13 00:00:00 01 41 15078 0.000710
4 2022-04-13 00:00:00 01 42 2077 0.0000978
5 2022-04-13 00:00:00 01 43 2554 0.000120
6 2022-04-13 00:00:00 01 44 29190 0.00137
7 2022-04-13 00:00:00 01 45 2065 0.0000972
8 2022-04-13 00:00:00 01 46 2054 0.0000967
9 2022-04-13 00:00:00 01 47 2625 0.000124
10 2022-04-13 00:00:00 01 48 146334 0.00689
# ... with 8,714 more rows
Use:
ylim("20220427", "20220420", "20220413")
I want to use the Prophet() function in R, but I cannot transform my column "YearWeek" to a as.Date() column.
I have a column "YearWeek" that stores values from 201401 up to 201937 i.e. starting in 2014 week 1 up to 2019 week 37.
I don't know how to declare this column as a date in the form yyyy-ww needed to use the Prophet() function.
Does anyone know how to do this?
Thank you in advance.
One solution could be to append a 01 to the end of your yyyy-ww formatted dates.
Data:
library(tidyverse)
df <- cross2(2014:2019, str_pad(1:52, width = 2, pad = 0)) %>%
map_df(set_names, c("year", "week")) %>%
transmute(date = paste(year, week, sep = "")) %>%
arrange(date)
head(df)
#> # A tibble: 6 x 1
#> date
#> <chr>
#> 1 201401
#> 2 201402
#> 3 201403
#> 4 201404
#> 5 201405
#> 6 201406
Now let's append the 01 and convert to date:
df %>%
mutate(date = paste(date, "01", sep = ""),
new_date = as.Date(date, "%Y%U%w"))
#> # A tibble: 312 x 2
#> date new_date
#> <chr> <date>
#> 1 20140101 2014-01-05
#> 2 20140201 2014-01-12
#> 3 20140301 2014-01-19
#> 4 20140401 2014-01-26
#> 5 20140501 2014-02-02
#> 6 20140601 2014-02-09
#> 7 20140701 2014-02-16
#> 8 20140801 2014-02-23
#> 9 20140901 2014-03-02
#> 10 20141001 2014-03-09
#> # ... with 302 more rows
Created on 2019-10-10 by the reprex package (v0.3.0)
More info about a numeric week of the year can be found here.
I have a time serie data from two consecutive years (2017-2018), from january to december of each year. Then I need to plot the data from sept-17 to april-18.
I could do it with a very hand-made code, however I realize it could be done very much straightforward way with the packages availabe today for managing dates on plots (packages "scales", "lubridate", etc.)
Can someone help me to simplify my work for doing the second plot?
I will really appreciate it.
suppressWarnings(suppressMessages(library("tidyverse", quietly = T)))
dat <- tibble(
date = seq(as.Date("2017-01-01"), as.Date("2018-12-31"), by=1),
var = rgamma(length(date), shape=2, scale=2)) %>%
mutate(year = lubridate::year(date),
month = lubridate::month(date),
julian = lubridate::yday(date))
dat
#> # A tibble: 730 x 5
#> date var year month julian
#> <date> <dbl> <dbl> <dbl> <dbl>
#> 1 2017-01-01 12.9 2017 1 1
#> 2 2017-01-02 6.69 2017 1 2
#> 3 2017-01-03 6.11 2017 1 3
#> 4 2017-01-04 1.68 2017 1 4
#> 5 2017-01-05 1.22 2017 1 5
#> 6 2017-01-06 10.2 2017 1 6
#> 7 2017-01-07 5.13 2017 1 7
#> 8 2017-01-08 4.61 2017 1 8
#> 9 2017-01-09 3.79 2017 1 9
#> 10 2017-01-10 1.11 2017 1 10
#> # … with 720 more rows
dat %>%
ggplot() +
geom_line(aes(julian, var, color = factor(month), linetype=factor(year)))
dat %>%
filter((year == 2017 & month %in% c("9","10", "11", "12"))|
(year == 2018 & month %in% c("1", "2", "3"))) %>%
mutate(julian_AWS = ifelse(julian>=244, julian-243, julian+123)) %>%
ggplot() +
geom_line(aes(julian_AWS, var, color = factor(month), linetype=factor(year)))+
scale_x_continuous(breaks = c(1,#S
31,#O
61,#N
91,#D
121,#E
151,#F
181),#M
labels = c("Sep", "Oct", "Nov", "Dec", "Jan", "Feb", "Mar"))+
theme(axis.text.x=element_text(hjust=-1))
Created on 2019-05-05 by the reprex package (v0.2.1)
I don't think you need to delve into the julian date formats. See if this gets you what you need:
dat %>%
filter(date >= '2017-09-01', date < '2018-04-01') %>%
ggplot() +
geom_line(aes(date, var, color = factor(month), linetype = factor(year))) +
scale_x_date(date_breaks = "1 month", date_labels = "%b") +
theme(axis.text.x = element_text(hjust = -1))
For more info on date label formats, see ?strftime
I'm having trouble figuring out how to use xblocks() to work. First, here's a small example from a much larger dataset:
data <- data.frame(
Date = sample(c("1993-07-05", "1993-07-05", "1993-07-05", "1993-08-30", "1993-08-30", "1993-08-30", "1993-08-30", "1993-09-04", "1993-09-04")),
Oxygen = sample(c("0.9", "0.4", "4.2", "5.6", "7.3", NA, "9.5", NA, "0.3")))
I then averaged values for each month using xts:
xtsAveragedata <- xts(Averagedata[-1], Averagedata[[1]])
xtsAverageMonthlyData <- apply.monthly(xtsAveragedata, FUN = mean)
Now, I linear interpolated my data:
Interpolateddata <- na.approx(xtsAverageMonthlyData)
I want to create a figure in which I use xblocks() or something similar to show the regions in my data where I used interpolation, something like this, which I found online:
How do I go about doing this for all values/automate for my entire dataset? There's no examples I could translate into something like this from the reference guide.
Thank you for your help. It is greatly appreciated.
So this doesn't use xts or zoo, but maybe this walkthrough will be helpful. I am using a slightly larger (and daily) dataset, but it should be reproducible:
library(tidyverse)
library(lubridate)
set.seed(4)
df <- tibble(
Date = seq.Date(ymd("1993-07-01"), by = "1 day", length.out = 100),
Oxygen = runif(100, 0, 10)
)
# Randomly assign 20 records to NA
df[sample(1:nrow(df), 20), "Oxygen"] <- NA
df_for_plot <- df %>%
arrange(Date) %>%
group_by(month(Date)) %>%
mutate(
is_na = is.na(Oxygen),
month_avg = mean(Oxygen, na.rm = TRUE),
oxygen_to_plot = if_else(is_na, month_avg, Oxygen)
)
df_for_plot
#> # A tibble: 100 x 6
#> # Groups: month(Date) [4]
#> Date Oxygen `month(Date)` is_na month_avg oxygen_to_plot
#> <date> <dbl> <dbl> <lgl> <dbl> <dbl>
#> 1 1993-07-01 5.86 7 FALSE 5.87 5.86
#> 2 1993-07-02 0.0895 7 FALSE 5.87 0.0895
#> 3 1993-07-03 2.94 7 FALSE 5.87 2.94
#> 4 1993-07-04 2.77 7 FALSE 5.87 2.77
#> 5 1993-07-05 8.14 7 FALSE 5.87 8.14
#> 6 1993-07-06 NA 7 TRUE 5.87 5.87
#> 7 1993-07-07 7.24 7 FALSE 5.87 7.24
#> 8 1993-07-08 9.06 7 FALSE 5.87 9.06
#> 9 1993-07-09 9.49 7 FALSE 5.87 9.49
#> 10 1993-07-10 0.731 7 FALSE 5.87 0.731
#> # ... with 90 more rows
# Plot the regular data, but for the geom_rect use only the filtered data where the is_na column is TRUE.
# Assuming you have daily data, you just set the xmax to be that Date + 1.
ggplot(df_for_plot, aes(x = Date, y = oxygen_to_plot)) +
geom_line() +
geom_rect(
data = df_for_plot %>% filter(is_na),
aes(xmin = Date, xmax = Date + 1, ymin = -Inf, ymax = +Inf), fill = "skyblue", alpha = 0.5
)
When I create my graph with geom line I get this graph
Here is my code
ggplot(Moisture_kurokawa, aes(x = Date))+ geom_line(aes(y = W5, colour = "W5"))+ geom_line(aes(y = W7, colour = "W7"))+ geom_line(aes(y = W9, colour = "W9"))+ geom_line(aes(y = W11, colour = "W11"))
Any help on how to get it smoothen or see all the data points?
My data file link.
You should take some time to read through some ggplot2 tutorials so you can let it do some of the work for you, especially if you do a bit of data rearranging.
You also need to have a proper date+time object for the resolution you want.
library(tidyverse)
Moisture_kurokawa <- read_csv("~/Data/Moisture kurokawa.csv")
mutate(Moisture_kurokawa,
timestamp = lubridate::mdy_hms(sprintf("%s %s", Date, Time))) %>%
select(-Date, -Time) %>%
gather(W, value, -timestamp) -> moisture_long
moisture_long
## # A tibble: 17,645 x 3
## timestamp W value
## <dttm> <chr> <dbl>
## 1 2017-06-24 00:00:00 W5 0.333
## 2 2017-06-24 00:30:00 W5 0.333
## 3 2017-06-24 01:00:00 W5 0.334
## 4 2017-06-24 01:30:00 W5 0.334
## 5 2017-06-24 02:00:00 W5 0.334
## 6 2017-06-24 02:30:00 W5 0.334
## 7 2017-06-24 03:00:00 W5 0.335
## 8 2017-06-24 03:30:00 W5 0.335
## 9 2017-06-24 04:00:00 W5 0.335
## 10 2017-06-24 04:30:00 W5 0.335
## # ... with 17,635 more rows
ggplot(moisture_long, aes(timestamp, value, group=W, color=W)) +
geom_line()
With your data in a better shape, you can even do:
ggplot(moisture_long, aes(timestamp, value, group=W, color=W)) +
geom_line() +
facet_wrap(~W)
Moisture_kurokawa <- read.table("Moisture kurokawa.csv", header=T, sep=",")
# Create a datetime object with as.POSIXct
Moisture_kurokawa$DateTime <- as.POSIXct(
paste0(Moisture_kurokawa$Date, Moisture_kurokawa$Time),
format="%m/%d/%Y %H:%M")
library(ggplot2)
ggplot(Moisture_kurokawa, aes(x = DateTime))+
geom_line(aes(y = W5, colour = "W5"))+
geom_line(aes(y = W7, colour = "W7"))+
geom_line(aes(y = W9, colour = "W9"))+
geom_line(aes(y = W11, colour = "W11"))