Error with ordering on facet_wrap geom_col

[hopnstop]

I'm using a facet_wrap combined with a geom_col, but the ordering isn't how I'd like it to appear - below is the graph as it looks now and you can clearly see the ordering in the top section is not correct:

Unfortunately, I've tried several things to fix this, but can't figure it out. Here's the code I'm using for the table as well as the data:

library(tidyverse)
library(dplyr)
library(ggimage)
library(ggrepel)
library(nflfastR)
library(magick)
library(cowplot)
library(showtext)
library(jsonlite)
library(curl)
library(rsvg)
library(ggsci)
library(ggridges)
library(ggthemes)
library(shadowtext)


font_add_google("Encode Sans Condensed", "encode", regular.wt = 400, bold.wt = 600)
font_add_google("Inconsolata", "incon")
showtext_auto()

ggplot(data = arizona,aes(x = pct_team_ep, y = reorder(player_name, rank), fill = Team))+
  geom_col(show.legend = FALSE, position = "identity")+
  geom_shadowtext(aes(x = pct_team_ep + 0.02, color = Team, label = player_name), typeface = "bold", hjust = 0, bg.color = 'white', family = "incon", size = 3.5, na.rm = T, bg.r = 0.2, show.legend = FALSE) +
  scale_x_continuous(labels = scales::percent_format(accuracy=1),
                     expand = expansion(mult = 0), limits = c(0,1),
                     name = "",
                     breaks = scales::pretty_breaks(5)) +
  scale_fill_manual(values = NFL_pri) +
  scale_color_manual(values = NFL_sec) +
  facet_wrap( ~ ep_type, ncol = 1, scales = "free") +
  theme_bw()+
  theme(
    axis.title.y = element_blank(),
    axis.text.y = element_blank(),
    axis.ticks.y = element_blank(),
    axis.ticks.x = element_line(size = 0.5),
    strip.background = element_rect(
      fill= "white", size=1, linetype="solid"),
    strip.text.x = element_text(
      size = 12, color = arizona$team_color, family = "encode", face = "bold"
    ))

Team	player_name	team_color	ep_type	pct_team_ep	rank
TB	B. Gabbert	#d50a0a	% Team Rushing Expected Fantasy Points	0.05803571	4
TB	C. Godwin	#d50a0a	% Team Receiving Expected Fantasy Points	0.22020287	14
TB	J. Watson	#d50a0a	% Team Receiving Expected Fantasy Points	0.05592323	3
TB	K. Vaughn	#d50a0a	% Team Receiving Expected Fantasy Points	0.05365854	2
TB	K. Vaughn	#d50a0a	% Team Rushing Expected Fantasy Points	0.11612903	9
TB	L. Fournette	#d50a0a	% Team Receiving Expected Fantasy Points	0.07182683	6
TB	L. Fournette	#d50a0a	% Team Rushing Expected Fantasy Points	0.31750339	15
TB	L. McCoy	#d50a0a	% Team Receiving Expected Fantasy Points	0.06256446	5
TB	M. Evans	#d50a0a	% Team Receiving Expected Fantasy Points	0.16322217	13
TB	O. Howard	#d50a0a	% Team Receiving Expected Fantasy Points	0.12724307	11
TB	R. Gronkowski	#d50a0a	% Team Receiving Expected Fantasy Points	0.15147957	12
TB	R. Jones	#d50a0a	% Team Receiving Expected Fantasy Points	0.09816815	7
TB	R. Jones	#d50a0a	% Team Rushing Expected Fantasy Points	0.57880435	16
TB	S. Miller	#d50a0a	% Team Receiving Expected Fantasy Points	0.11648661	10
TB	T. Brady	#d50a0a	% Team Rushing Expected Fantasy Points	0.11413043	8
TB	T. Johnson	#d50a0a	% Team Receiving Expected Fantasy Points	0.05351488	1

Any suggestions on what to do? Let me know and thanks in advance!

[ZykeZero]

you can try factor reorder from forcats
when you share data you might want to share it in a format that can be copied out for others to test. usually using dput(data) / dput(arizona) in your case.

library(forcats)

arizona %>%
  mutate(player_name = fct_reorder(player_name, rank)) %>%
  ggplot(...)

[hopnstop]

Thanks, I'll be sure to do that in the future! Unfortunately, the fct_reorder didn't work for me. Any other solutions you can think of?

[andresrcs]

You can use reorder_within() from the tidytext package

library(ggplot2)
library(tidytext)

arizona <- data.frame(
  stringsAsFactors = FALSE,
              Team = c("B.","C.","J.","K.","K.",
                       "L.","L.","L.","M.","O.","R.","R.","R.","S.","T.",
                       "T."),
       player_name = c("Gabbert","Godwin","Watson",
                       "Vaughn","Vaughn","Fournette","Fournette","McCoy",
                       "Evans","Howard","Gronkowski","Jones","Jones","Miller",
                       "Brady","Johnson"),
        team_color = c("#d50a0a","#d50a0a","#d50a0a",
                       "#d50a0a","#d50a0a","#d50a0a","#d50a0a","#d50a0a",
                       "#d50a0a","#d50a0a","#d50a0a","#d50a0a","#d50a0a",
                       "#d50a0a","#d50a0a","#d50a0a"),
           ep_type = c("% Team Rushing Expected Fantasy Points","% Team Receiving Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points",
                       "% Team Rushing Expected Fantasy Points","% Team Receiving Expected Fantasy Points",
                       "% Team Rushing Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points","% Team Receiving Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points",
                       "% Team Rushing Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points","% Team Rushing Expected Fantasy Points",
                       "% Team Receiving Expected Fantasy Points"),
       pct_team_ep = c(0.05803571,0.22020287,
                       0.05592323,0.05365854,0.11612903,0.07182683,0.31750339,
                       0.06256446,0.16322217,0.12724307,0.15147957,0.09816815,
                       0.57880435,0.11648661,0.11413043,0.05351488),
              rank = c(4, 14, 3, 2, 9, 6, 15, 5, 13, 11, 12, 7, 16, 10, 8, 1)
)

ggplot(data = arizona, aes(x = pct_team_ep,
                           y = reorder_within(player_name, rank, ep_type),
                           fill = Team)) +
    geom_col(show.legend = FALSE) +
    scale_x_continuous(labels = scales::percent_format(accuracy=1),
                       expand = expansion(mult = 0), limits = c(0,1),
                       name = "",
                       breaks = scales::pretty_breaks(5)) +
    scale_y_reordered() +
    facet_wrap( ~ ep_type, ncol = 1, scales = "free")

^{Created on 2020-10-29 by the reprex package (v0.3.0.9001)}

Note: Next time please post a proper REPRoducible EXample (reprex) illustrating your issue as explained on this link.