r - 使用 facet_wrap 时 ggplot 重新排序更改

Question

我正在尝试重新排序我的数据，然后使用 facet_wrap 绘制不同的结果。

当我使用以下内容时:

df %>% 
  ggplot(aes(x = reorder(Feature, Gain), y = Gain)) +
  geom_bar(stat = "identity") +
  coord_flip()

绘图“似乎”按预期工作，即条形图从最高到最低排序。但是，当我添加代码的 facet_wrap 部分时，绘图会发生变化。

df %>% 
  ggplot(aes(x = reorder(Feature, Gain), y = Gain)) +
  geom_bar(stat = "identity") +
  coord_flip() +
  facet_wrap(~Model_Name)

我知道原因是我正在根据整个数据集重新排序数据，然后当我应用这些方面时，这就是它们发生变化的原因。我如何根据每个方面使用重新排序，以便每个方面总是在顶部有最高的结果，在底部有最低的结果？

数据:

df <- structure(list(Model_Name = c("1", "1", "1", "1", "1", "1", "1", 
"1", "1", "1", "1", "2", "2", "2", "2", "2", "2", "2", "2", "2", 
"2", "2", "3", "3", "3", "3", "3", "3", "3", "3", "3", "3", "3", 
"4", "4", "4", "4", "4", "4", "4", "4", "4", "4", "4"), Feature = c("EBIT.FinExp", 
"DailySALES.EBIT", "logTA", "TL.TA", "CF.NCL", "TL.EQ", "CL.FinExp", 
"CA.CL", "CA.TA", "EQ.Turnover", "SALES.WC", "logTA", "EBIT.FinExp", 
"TL.TA", "EQ.Turnover", "CL.FinExp", "CA.TA", "TL.EQ", "CF.NCL", 
"CA.CL", "DailySALES.EBIT", "SALES.WC", "logTA", "TL.TA", "EBIT.FinExp", 
"TL.EQ", "CA.TA", "CL.FinExp", "EQ.Turnover", "CA.CL", "SALES.WC", 
"DailySALES.EBIT", "CF.NCL", "logTA", "TL.TA", "TL.EQ", "CA.TA", 
"CL.FinExp", "EQ.Turnover", "SALES.WC", "EBIT.FinExp", "CA.CL", 
"DailySALES.EBIT", "CF.NCL"), Gain = c(0.319661022583932, 0.1459294841629, 
0.142850179507091, 0.132765644191462, 0.0810216258897842, 0.0427415667709611, 
0.0398929568321689, 0.038977683436756, 0.0235407288506038, 0.0177791651224692, 
0.0148399426518705, 0.314495967867966, 0.194685182626366, 0.140733861259937, 
0.0882380538977587, 0.0574929508879213, 0.0465452034142143, 0.0409126137922541, 
0.0356003145682383, 0.0300757198164929, 0.029110471062585, 0.0221096608062664, 
0.397020881823162, 0.143886886033381, 0.0915945975581377, 0.0699399814849332, 
0.0633989152531246, 0.0608103962252964, 0.0512060957146738, 0.039355302553383, 
0.0342297915015821, 0.0255241325884949, 0.0230330192638316, 0.38056229652222, 
0.133231017163002, 0.0941844467535513, 0.0788600251129408, 0.0749519130106986, 
0.0521907558198196, 0.0498475034974015, 0.0433895506999798, 0.0428104836292833, 
0.0273526930263904, 0.0226193147647128), Cover = c(0.121943691387452, 
0.0816186622203503, 0.234722543977359, 0.171634330582589, 0.0892304037319097, 
0.0315468239516433, 0.0963968265880753, 0.0607290547248047, 0.0452337706760221, 
0.0279521362780545, 0.0389917558817403, 0.227632942870328, 0.0878012612621709, 
0.132093360664219, 0.0904386168640914, 0.109620838100331, 0.0915521745157415, 
0.0239824724837257, 0.0573896745122457, 0.0601712810342116, 0.0639922580136253, 
0.0553251196793097, 0.231725278554576, 0.103655920813926, 0.0814217672437898, 
0.0491388474456846, 0.118941515640324, 0.0964283304134241, 0.0947214507718667, 
0.0741550093313701, 0.0558597882883057, 0.0425784549213668, 0.0513736365753665, 
0.201160082759667, 0.0826529411502635, 0.0493477928111075, 0.111213779712931, 
0.110738868440133, 0.102709355008914, 0.0534211062162858, 0.101876672415992, 
0.067425714362272, 0.0636683097957719, 0.0557853773266621), Frequency = c(0.104681978798587, 
0.0940812720848057, 0.152385159010601, 0.125883392226148, 0.0936395759717314, 
0.0415194346289753, 0.0954063604240283, 0.0874558303886926, 0.075530035335689, 
0.0636042402826855, 0.0658127208480565, 0.141749246014649, 0.0896165445928479, 
0.112882378285222, 0.0934941835415769, 0.0947867298578199, 0.100387763894873, 
0.042654028436019, 0.0779836277466609, 0.0822921154674709, 0.0861697544161999, 
0.0779836277466609, 0.157513835674755, 0.0906768837803321, 0.0936568752660707, 
0.0510855683269476, 0.108982545764155, 0.102596849723287, 0.0932311621966794, 
0.0872711792252022, 0.0825883354618987, 0.0723712217965091, 0.0600255427841635, 
0.121157323688969, 0.0822784810126582, 0.0592224231464738, 0.117088607594937, 
0.0981012658227848, 0.099005424954792, 0.0981012658227848, 0.0958408679927667, 
0.0840867992766727, 0.0782097649186257, 0.0669077757685353)), row.names = c(NA, 
-44L), class = c("data.table", "data.frame"))

Answer 1

Github 上 David Robinson 的个人 R 包中有一个名为 reorder_within 的好函数，drlib .它旨在完全满足您的需求。

remotes::install_github("dgrtwo/drlib")
df %>% 
    ggplot(aes(x = drlib::reorder_within(Feature, Gain, within = Model_Name), y = Gain)) +
    geom_bar(stat = "identity") +
    coord_flip() +
    facet_wrap(~Model_Name, scales = "free")

您可以使用下面的 stats::reorder 自行实现他的代码，但希望在应有的地方给予信任。

df %>% 
    ggplot(aes(x = stats::reorder(paste(Feature, Model_Name, sep = "__"), Gain), y = Gain)) +
    geom_bar(stat = "identity") +
    coord_flip() +
    facet_wrap(~Model_Name, scales = "free")