Plot Timeseries Data where the hourly variations are values in different columns-CodePudding

I have a DataFrame as shown below. The Columns 0:00, 1:00.. till 24:00 represents the hours. How can I plot the hourly variation of both Items 'foo' and 'bar' for all the dates?

Date        Item  0:00   1:00  2:00  3:00  4:00  5:00  6:00 

1/1/2022    foo   2      3    4       1      5     1.5   2.5 
1/2/2022    foo   1.5    1    3       2      2.5   6     4
1/3/2022    foo   1      2    3       1      2     5     4
1/4/2022    foo   2      1    3       4      1     2.4   3
1/1/2022    bar   3      1    0       1.5    5     1.5   2.5
1/2/2022    bar   2.5    4    3       1      2.5   0     1
1/3/2022    bar   1      2    1       1      2     1.5   4
1/4/2022    bar   2      1    3       2      1     2.5   3

I tried the following

g = sns.FacetGrid(df,row='Item',col='Date')
g.map(sns.#someplot,) # within map not sure what plot should I use and how to represent x axis as the columns

CodePudding user response：

If need bars with separate plots for Items use:

df1 = df.melt(['Date','Item'])
g = sns.FacetGrid(df1, row='Item', col="Date")
g.map_dataframe(sns.barplot, x="variable", y="value")

Or if need both Items in one graph use:

df1 = df.melt(['Date','Item'])
g = sns.FacetGrid(df1, col="Date")
g.map_dataframe(sns.barplot, x="variable", y="value", hue="Item")

CodePudding user response：

With R, you have to pivot to long format, particularly when using ggplot2.

library(tidyverse) 

df %>%
  pivot_longer(-c(Date, Item), names_to = "time", values_to = "value") %>%
  mutate(Date = as.Date(Date, "%m/%d/%Y")) %>%
  mutate(time = as.POSIXct(paste(ok$Date, ok$time), format="%Y-%m-%d %H:%M")) %>%
  ggplot(aes(x = time, y = value, colour = Item, group = Item))  
  geom_line()

*Note: Since it is just a sample of the data, there are time gaps, hence the sharp angles.

Or if you want them separate, then you can facet the plot (also using bar for another alternative).

df %>%
  pivot_longer(-c(Date, Item), names_to = "time", values_to = "value") %>%
  mutate(Date = as.Date(Date, "%m/%d/%Y")) %>%
  mutate(time = as.POSIXct(paste(ok$Date, ok$time), format="%Y-%m-%d %H:%M")) %>%
  ggplot(aes(x = time, y = value))  
  geom_bar(stat = "identity", aes(colour = Item, fill = Item))  
  facet_grid(~Item, scale='free_x')

Data

df <- structure(list(Date = c("1/1/2022", "1/2/2022", "1/3/2022", "1/4/2022", 
"1/1/2022", "1/2/2022", "1/3/2022", "1/4/2022"), Item = c("foo", 
"foo", "foo", "foo", "bar", "bar", "bar", "bar"), `0:00` = c(2, 
1.5, 1, 2, 3, 2.5, 1, 2), `1:00` = c(3L, 1L, 2L, 1L, 1L, 4L, 
2L, 1L), `2:00` = c(4L, 3L, 3L, 3L, 0L, 3L, 1L, 3L), `3:00` = c(1, 
2, 1, 4, 1.5, 1, 1, 2), `4:00` = c(5, 2.5, 2, 1, 5, 2.5, 2, 1
), `5:00` = c(1.5, 6, 5, 2.4, 1.5, 0, 1.5, 2.5), `6:00` = c(2.5, 
4, 4, 3, 2.5, 1, 4, 3)), class = "data.frame", row.names = c(NA, 
-8L))

CodePudding user response：

In R just use matplot, no need for data wrangling beforehand.

par(mfrow=c(1, 2)); by(df[3:9], df$Item, \(x) matplot(t(x), type='l'))

Or, more publishable:

par(mfrow=c(1, 2))
by(df, df$Item, \(x) {
  matplot(t(x[1:4, 3:9]), type='l', main=el(x$Item), xaxt='n', xlab='t', ylab='y')
  axis(1, axTicks(1), colnames(df)[3:9])
  legend('topleft', leg=x$Date, col=seq_len(nrow(x)), lty=seq_len(nrow(x)), 
         cex=.8, ncol=2)
  })

Data:

df <- read.table(text="
                 Date Item 0:00 1:00 2:00 3:00 4:00 5:00 6:00
1 1/1/2022  foo  2.0    3    4  1.0  5.0  1.5  2.5
2 1/2/2022  foo  1.5    1    3  2.0  2.5  6.0  4.0
3 1/3/2022  foo  1.0    2    3  1.0  2.0  5.0  4.0
4 1/4/2022  foo  2.0    1    3  4.0  1.0  2.4  3.0
5 1/1/2022  bar  3.0    1    0  1.5  5.0  1.5  2.5
6 1/2/2022  bar  2.5    4    3  1.0  2.5  0.0  1.0
7 1/3/2022  bar  1.0    2    1  1.0  2.0  1.5  4.0
8 1/4/2022  bar  2.0    1    3  2.0  1.0  2.5  3.0
           ", check.names=F)