rggplot2legendaesthetics
Can someone help me properly set up my ggplot legend with multiple defined aesthetics?
I would love to have a clearer legend in my ggplot (below). Specifically, the point shapes overlapping with line geometries is not clear. It would be nice to have one legend for the predictive data (linetypes, ribbon fills, color) and another for the raw data (point shapes, color).
After playing around I managed to separate the two groups, but I can't find a way to have group colors applied to both legends. Furthermore, it seems that even when I group linetype and fill together using the same labs() name, there is still a grey fill on the point shape legend. This is a lot of information but the takeaway is after much playing around, I can't get the legend to behave.
I have provided my code and datasets; maybe someone could help me out?
Here's my ggplot:
and here's the code that generated it:
ggplot(predict.df, aes(x = x, y = predicted, colour = group, shape = group, linetype = group)) +
stat_smooth(method = "lm", formula = y ~ x) +
geom_ribbon(aes(ymin = conf.low, ymax = conf.high, fill = group), alpha = 0.1) +
geom_jitter(data = raw, mapping = aes(x = x, y = response), width = 8, height = 0) +
scale_x_continuous(name = "Sample date", breaks = c(0, 35, 70, 105), labels = c("June 26", "July 31", "Sept 4", "Oct 9")) +
scale_y_continuous(name = "Viral load (virus/\u00b5g DNA)", breaks = c(0, 1, 2, 3), labels = c(1, 10, 100, 1000)) +
theme_pubr(legend = "right") +
labs(color = "Locus", shape = "Locus", fill = "Locus", linetype = "Locus") +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2")
Here are the two plugged in data frames:
> dput(predict.df)
structure(list(x = c(0, 0, 0, 35, 35, 35, 70, 70, 70, 105, 105,
105), predicted = c(1.76102123590214, 0.37702177715769, 0.502111657963439,
2.16765850174448, 1.13876128504506, 0.72723857556493, 2.57429576758682,
1.90050079293243, 0.95236549316642, 2.98093303342917, 2.66224030081979,
1.17749241076791), std.error = c(0.210212347913819, 0.167399741123415,
0.183879503224061, 0.145514809621925, 0.106824868721082, 0.108685998504236,
0.121276566028898, 0.176399586908295, 0.18802919804674, 0.15744703287395,
0.295192239702551, 0.323169724522022), conf.low = c(1.34901260488545,
0.0489243135344682, 0.141714454149163, 1.88245471566831, 0.929388389698517,
0.514217932872852, 2.33659806600149, 1.55476395570442, 0.58383503695286,
2.67234251952353, 2.08367414248708, 0.544091389811016), conf.high = c(2.17302986691883,
0.705119240780912, 0.862508861777715, 2.45286228782065, 1.3481341803916,
0.940259218257007, 2.81199346917216, 2.24623763016043, 1.32089594937998,
3.2895235473348, 3.24080645915251, 1.8108934317248), group = structure(c(1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L), .Label = c("Basal",
"Medial", "Distal"), class = "factor")), row.names = c(NA, -12L
), class = c("ggeffects", "data.frame"), legend.labels = c("Basal",
"Medial", "Distal"), x.is.factor = "0", continuous.group = FALSE, rawdata = structure(list(
response = c(2.23667815433731, 0.585725623985699, 0, 2.63585811315419,
0, 0, 2.33374975605124, 0, 0.884942436609621, 1.16286703780418,
1.06709808539313, 0, 0.890758164061038, 0, 0.639404117570214,
2.47790759499121, 0, 0, 2.14524220682092, 0, 0, 1.9740589910418,
0, 0, 0, 1.25516666996646, 2.66751722923715, 0, 0, 1.92521830152371,
0.903597044237988, 0.945099129999455, 2.6442228874261, 2.58142029334097,
0, 2.32153495885343, 1.3801896215756, 1.02489379026788, 2.05743286900869,
1.11074999614049, 0.860338006570994, 1.55767122072119, 0.989126434016326,
1.21554496604287, 2.14770759585995, 1.92492507723203, 1.86644529881602,
2.40375462277453, 1.34200834355252, 1.14456300882461, 2.25413032451885,
1.05881158464742, 0.880851309461926, 2.72704842428525, 1.01414176946023,
1.13102772933728, 2.78511309801747, 2.0899051114394, 1.02716459664481,
1.90825217816864, 0.791001136920094, 2.27946840305546, 2.48278954016739,
0.940828612330139, 2.20441891726465, 1.95118485766815, 1.01752434498353,
1.22992186490351, 1.18351341124408, 1.17491390968833, 1.54441258580579,
1.24401572811876, 0.878566226876958, 2.35327316999556, 2.13603849721832,
2.29146399657572, 2.41044404137815, 2.21254419548501, 0.826472233221594,
2.20421121140994, 1.57587890647473, 0.897437484440932, 2.34225098899079,
2.38136059256142, 0.972352525321007, 1.6028414421383, 2.20134373839759,
0.810820367137012, 2.77026731894286, 2.29241135584685, 0.57988940596851,
3.71170452590752, 3.4954877455041, 3.18504082301818, 3.67423352412794,
2.98414705180641, 0.141639163861031, 2.41217775364379, 0.51123419122885,
0.476456687670469, 3.36887472776869, 3.02081502150107, 0.513445669923521,
3.63514892267612, 3.51925820816477, 3.40746252696249, 3.27998529166281,
0.623505586942313, 0.798307744015814, 3.27752745673065, 3.18652456493964,
0.184008911262831, 3.18983391030149, 2.77631390383718, 0.395623039048979,
3.46567596502934, 3.02900875924267, 0.433385505287688), x = c(0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 35, 35, 35, 35, 35, 35, 35, 35,
35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
35, 35, 35, 35, 35, 35, 35, 70, 70, 70, 70, 70, 70, 70, 70,
70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70,
70, 70, 70, 70, 70, 70, 105, 105, 105, 105, 105, 105, 105,
105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105,
105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105), group = structure(c(1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L), .Label = c("Basal",
"Medial", "Distal"), class = "factor")), class = "data.frame", row.names = c(NA,
-118L)), title = "Predicted values of log.Virus.Particles", x.title = "Date", y.title = "log.Virus.Particles", legend.title = "Locus", constant.values = list(
Study.ID = "0 (population-level)"), terms = c("Date", "Locus"
), original.terms = c("Date", "Locus"), at.list = list(Date = c(0,
35, 70, 105), Locus = c("Basal", "Medial", "Distal")), prediction.interval = FALSE, ci.lvl = 0.95, family = "gaussian", link = "identity", logistic = "0", is.trial = "0", fitfun = "lm", model.name = "model1")
and
> dput(raw)
structure(list(response = c(2.23667815433731, 0.585725623985699,
0, 2.63585811315419, 0, 0, 2.33374975605124, 0, 0.884942436609621,
1.16286703780418, 1.06709808539313, 0, 0.890758164061038, 0,
0.639404117570214, 2.47790759499121, 0, 0, 2.14524220682092,
0, 0, 1.9740589910418, 0, 0, 0, 1.25516666996646, 2.66751722923715,
0, 0, 1.92521830152371, 0.903597044237988, 0.945099129999455,
2.6442228874261, 2.58142029334097, 0, 2.32153495885343, 1.3801896215756,
1.02489379026788, 2.05743286900869, 1.11074999614049, 0.860338006570994,
1.55767122072119, 0.989126434016326, 1.21554496604287, 2.14770759585995,
1.92492507723203, 1.86644529881602, 2.40375462277453, 1.34200834355252,
1.14456300882461, 2.25413032451885, 1.05881158464742, 0.880851309461926,
2.72704842428525, 1.01414176946023, 1.13102772933728, 2.78511309801747,
2.0899051114394, 1.02716459664481, 1.90825217816864, 0.791001136920094,
2.27946840305546, 2.48278954016739, 0.940828612330139, 2.20441891726465,
1.95118485766815, 1.01752434498353, 1.22992186490351, 1.18351341124408,
1.17491390968833, 1.54441258580579, 1.24401572811876, 0.878566226876958,
2.35327316999556, 2.13603849721832, 2.29146399657572, 2.41044404137815,
2.21254419548501, 0.826472233221594, 2.20421121140994, 1.57587890647473,
0.897437484440932, 2.34225098899079, 2.38136059256142, 0.972352525321007,
1.6028414421383, 2.20134373839759, 0.810820367137012, 2.77026731894286,
2.29241135584685, 0.57988940596851, 3.71170452590752, 3.4954877455041,
3.18504082301818, 3.67423352412794, 2.98414705180641, 0.141639163861031,
2.41217775364379, 0.51123419122885, 0.476456687670469, 3.36887472776869,
3.02081502150107, 0.513445669923521, 3.63514892267612, 3.51925820816477,
3.40746252696249, 3.27998529166281, 0.623505586942313, 0.798307744015814,
3.27752745673065, 3.18652456493964, 0.184008911262831, 3.18983391030149,
2.77631390383718, 0.395623039048979, 3.46567596502934, 3.02900875924267,
0.433385505287688), x = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 35, 35, 35,
35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35,
35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 35, 70, 70, 70, 70, 70,
70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70, 70,
70, 70, 70, 70, 70, 70, 70, 70, 105, 105, 105, 105, 105, 105,
105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105,
105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105), group = structure(c(1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 2L, 3L, 1L, 2L, 3L, 1L,
2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L,
3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L,
1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L, 2L, 3L, 1L,
2L, 3L, 1L, 2L, 3L), .Label = c("Basal", "Medial", "Distal"), class = "factor")), class = "data.frame", row.names = c(NA,
-118L))
Edit 1, addressing MrFlick's comment
I tried the suggestions at the link you provided:
## Model with predictive slopes
ggplot(predict.df, aes(x = x, y = predicted, colour = group, shape = group, linetype = group)) +
stat_smooth(method = "lm", formula = y ~ x) +
geom_ribbon(aes(ymin = conf.low, ymax = conf.high, fill = group), alpha = 0.1) +
geom_jitter(data = raw, mapping = aes(x = x, y = response), width = 8, height = 0) +
scale_x_continuous(name = "Sample date", breaks = c(0, 35, 70, 105), labels = c("June 26", "July 31", "Sept 4", "Oct 9")) +
scale_y_continuous(name = "Viral load (virus/\u00b5g DNA)", breaks = c(0, 1, 2, 3), labels = c(1, 10, 100, 1000)) +
theme_pubr(legend = "right") +
scale_color_brewer(name = "Raw data", palette = "Dark2") +
scale_fill_brewer(name = "Prediction", palette = "Dark2") +
scale_linetype_discrete(name = "Prediction") +
scale_shape_discrete(name = "Raw data")
And ended up with:
This is more or less where I end up no matter what I try.
Solution
I think you can get what you want by altering the guides:
ggplot(predict.df, aes(x = x, y = predicted, colour = group,
shape = group, linetype = group)) +
stat_smooth(method = "lm", formula = y ~ x) +
geom_ribbon(aes(ymin = conf.low, ymax = conf.high, fill = group), alpha = 0.1) +
geom_jitter(data = raw, mapping = aes(x = x, y = response), width = 8, height = 0) +
scale_x_continuous(name = "Sample date", breaks = c(0, 35, 70, 105),
labels = c("June 26", "July 31", "Sept 4", "Oct 9")) +
scale_y_continuous(name = "Viral load (virus/\u00b5g DNA)",
breaks = c(0, 1, 2, 3),
labels = c(1, 10, 100, 1000)) +
theme_pubr(legend = "right") +
labs(color = "Raw", shape = "Raw", fill = "Locus", linetype = "Locus") +
scale_color_brewer(palette = "Dark2") +
scale_fill_brewer(palette = "Dark2") +
guides(fill = guide_legend(name = "Locus",
override.aes = list(
fill = RColorBrewer::brewer.pal(3, "Dark2"),
alpha = 0.1)),
linetype = guide_legend(name = "Locus",
override.aes = list(
color = RColorBrewer::brewer.pal(3, "Dark2"))),
shape = guide_legend(name = "Raw", override.aes = list(size = 3)),
color = guide_legend(name = "Raw",
override.aes = list(fill = NA,
linetype = 0)))
- R: attach within function
- Is it possible to get the index status (Index Coverage report in new version) from Google Search Console via the API
- Grouping string together and sum it together
- Association Rules in R - Trying to understand two similar datasets but one has rules but not the other
- How to draw vertical curly braces in a ggplot
- how to convert utf-8 to ascii for a list of data frames
- Insert appropriate number of rows into data frame based on date differences
- Remove an element of a list by name
- Querying API for every unique resource in a package using R
- Create a column in a dataframe based on comparisons of multiple columns in the original and a second dataframe
- calendR - "Error in if (special.days != "weekend" { : the condition has length > 1" for some datasets but not others
- Trying to create continent map for South America but geom_polygon produces zigzag fill
- Replacing NA cells in specific layers of a raster based on several conditions
- Extrapolating sequence of numbers in R
- How do I install a package that has been archived from CRAN?
- Calculating percent growth in time series data
- R Analysing data with different time stamps
- Imputing date in time series dataframe
- How to create an interactive map using echarty package in R?
- how to find unique characters both in forward and backward order in R
- Why the singular gradient error for this bi-exponential model
- Get the length of the list stored in a column of data.frame
- How can I output multiple flextable objects in R Markdown
- Conda environment loading a different R version than what I installed
- How to round a data.frame in R that contains some character variables?
- Prevent square brackets being turned into \link in conversion from ROxygen to Rd
- How to sort letters in a string?
- Find the code used to run one dynamic branch from id with targets
- How to deal with a number as column name, using an object
- Match the text formatting on a forest plot