I'm trying to create a graph like this:

I'm using ggplot2 function geom_linerange and geom_point to do this. In my graph I would like to put the year variable on the y axis, the coefficient on the x axis and the confidence interval with the geom_linerange but I am getting this error:

Error: Invalid input: date_trans works with objects of class Date only

If I remove the geom_linerange, the code works.

My code:

library(tidyverse)

results %>%
  group_by(Year) %>%
  ggplot(aes(x = Coefficient, y = Year, color = Variable))  
  geom_point()  
  geom_linerange(aes(x = Variable, ymin = conf.low_95, ymax = conf.high_95))

My dput:

structure(list(Variable = c("Model 1", "Model 2", "Model 3", 
"Model 4", "Model 5", "Model 6", "Model 1", "Model 2", "Model 3", 
"Model 4", "Model 5", "Model 6", "Model 1", "Model 2", "Model 3", 
"Model 4", "Model 5", "Model 6", "Model 1", "Model 2", "Model 3", 
"Model 4", "Model 5", "Model 6", "Model 1", "Model 2", "Model 3", 
"Model 4", "Model 5", "Model 6", "Model 1", "Model 2", "Model 3", 
"Model 4", "Model 5", "Model 6"), Coefficient = c(-0.337563422687641, 
-0.339870358919705, -0.174601667810286, -0.822153207427548, 0.0984725028254042, 
-0.302360726656967, -0.0144099316467553, 0.394261621304049, 0.408314375266859, 
-0.0437332910230634, 0.197918449352892, 0.211892879792213, -0.333777406931884, 
-0.211101666108092, -0.191240973701465, -0.200312029595054, 0.271332017169336, 
0.0198772428073128, -0.271118540022818, -0.296438420884506, -0.1959350798131, 
-0.604938094557287, -0.347759168324679, -0.565499762731695, 0.04352845127211, 
0.00464668179125901, 0.0618870586543521, 0.0154152406576902, 
-0.0882820982782311, -0.300512132812594, 0.354012886011799, 0.204557246817836, 
0.239302395932748, 0.419316974288796, -0.196096726024389, -0.0990227142573055
), conf.low_95 = c(-0.659882600540782, -0.652153038033736, -0.592355494245245, 
-1.15887028037482, -0.0957338272150386, -0.653747543529451, -0.360778831131537, 
-0.0547889861473285, -0.0813328193407771, -0.362272084151103, 
-0.00614433448580284, -0.0851580894901174, -0.690820142512398, 
-0.591538313876884, -0.665799398871222, -0.524239633810207, -0.00729694825470084, 
-0.312432149478429, -0.731712380556692, -0.742564839785077, -0.644569229627752, 
-1.00362824114186, -0.646168503067745, -0.872199479720429, -0.295291773790078, 
-0.366647571941097, -0.278411009308147, -0.308350900982826, -0.328643095941367, 
-0.598339525727505, -0.161377035034601, -0.674101088034365, -0.360447691535622, 
-0.0792532689562487, -0.541552105303034, -0.668221935319178), 
    conf.high_95 = c(-0.0152442448344994, -0.0275876798056746, 
    0.243152158624673, -0.485436134480278, 0.292678832865847, 
    0.0490260902155157, 0.331958967838026, 0.843312228755427, 
    0.897961569874495, 0.274805502104976, 0.401981233191587, 
    0.508943849074543, 0.0232653286486298, 0.1693349816607, 0.283317451468292, 
    0.1236155746201, 0.549960982593372, 0.352186635093054, 0.189475300511056, 
    0.149687998016065, 0.252699070001552, -0.206247947972715, 
    -0.049349833581613, -0.258800045742961, 0.382348676334298, 
    0.375940935523615, 0.402185126616851, 0.339181382298207, 
    0.152078899384905, -0.00268473989768353, 0.869402807058199, 
    1.08321558167004, 0.839052483401118, 0.917887217533841, 0.149358653254257, 
    0.470176506804567), Year = structure(c(17897, 17897, 17897, 
    17897, 17897, 17897, 18262, 18262, 18262, 18262, 18262, 18262, 
    18628, 18628, 18628, 18628, 18628, 18628, 18993, 18993, 18993, 
    18993, 18993, 18993, 19358, 19358, 19358, 19358, 19358, 19358, 
    19723, 19723, 19723, 19723, 19723, 19723), class = "Date")), row.names = c(NA, 
-36L), class = c("tbl_df", "tbl", "data.frame"))

CodePudding user response：

I think there are two issues:

1. I'm inferring that Coefficient and the conf.* variables should be on the same axis, but you are plotting x=Coefficient and then y*=conf* in the linerange geom. I suspect they should both be on the y-axis.

2. Your main mapping has (after swapping for the above) x=Year, but your linerange geom maps x=Variable, a categorical variable. Since you include color=Variable in the first mapping, I'll infer that that should be inherited in your linerange, so I'll remove the x= mapping and let it inherit.

In the end, this is a reasonable start:

results %>%
  group_by(Year) %>%
  ggplot(aes(y = Coefficient, x = Year, color = Variable))  
  geom_point()  
  geom_linerange(aes(ymin = conf.low_95, ymax = conf.high_95))

Side note ... it might be useful to break those lineranges apart a little. Here's a suggestion, though it is a bit fragile in that it relies on "120 days" as the dodging width, and this will be sensitive to the range of data years.

results %>%
  group_by(Year) %>%
  ggplot(aes(y = Coefficient, x = Year, color = Variable))  
  geom_point(position=position_dodge(width=120))  
  geom_linerange(aes(ymin = conf.low_95, ymax = conf.high_95), position=position_dodge(width=120))

This obviously won't work if the date samples are more closely-spaced. Just a thought :-)