Jarad
Reputation: 18953
Reorder R GGPlot Heatmap Layout While Preserving Correct Y-axis Labels
I am plotting a cohort analysis with R. I have a cohort triangle plot going but I want to flip the y-axis to sort the labels with "2009-01" at the top (the longest row) and "2010-03" at the bottom (shortest row). Doing so will flip the triangle from ◺ to ◸ which is what I want.
Data for a recreatable example:
melted <- read.table(text=" CohortPeriod variable value
1 1 2009-01 1.00000000
2 2 2009-01 0.36363636
3 3 2009-01 0.45454545
4 4 2009-01 0.40909091
5 5 2009-01 0.45454545
6 6 2009-01 0.36363636
7 7 2009-01 0.36363636
8 8 2009-01 0.31818182
9 9 2009-01 0.31818182
10 10 2009-01 0.31818182
11 11 2009-01 0.31818182
12 12 2009-01 0.36363636
13 13 2009-01 0.50000000
14 14 2009-01 0.31818182
15 15 2009-01 0.27272727
16 1 2009-02 1.00000000
17 2 2009-02 0.20000000
18 3 2009-02 0.33333333
19 4 2009-02 0.06666667
20 5 2009-02 0.26666667
21 6 2009-02 0.26666667
22 7 2009-02 0.26666667
23 8 2009-02 0.33333333
24 9 2009-02 0.33333333
25 10 2009-02 0.26666667
26 11 2009-02 0.20000000
27 12 2009-02 0.20000000
28 13 2009-02 0.33333333
29 14 2009-02 NA
30 15 2009-02 NA
31 1 2009-03 1.00000000
32 2 2009-03 0.30769231
33 3 2009-03 0.38461538
34 4 2009-03 0.30769231
35 5 2009-03 0.07692308
36 6 2009-03 0.15384615
37 7 2009-03 0.15384615
38 8 2009-03 0.23076923
39 9 2009-03 0.15384615
40 10 2009-03 0.07692308
41 11 2009-03 0.23076923
42 12 2009-03 0.15384615
43 13 2009-03 0.07692308
44 14 2009-03 NA
45 15 2009-03 NA
46 1 2009-04 1.00000000
47 2 2009-04 0.33333333
48 3 2009-04 0.25641026
49 4 2009-04 0.33333333
50 5 2009-04 0.15384615
51 6 2009-04 0.17948718
52 7 2009-04 0.10256410
53 8 2009-04 0.15384615
54 9 2009-04 0.05128205
55 10 2009-04 0.10256410
56 11 2009-04 0.07692308
57 12 2009-04 0.05128205
58 13 2009-04 NA
59 14 2009-04 NA
60 15 2009-04 NA
61 1 2009-05 1.00000000
62 2 2009-05 0.26000000
63 3 2009-05 0.24000000
64 4 2009-05 0.10000000
65 5 2009-05 0.08000000
66 6 2009-05 0.12000000
67 7 2009-05 0.06000000
68 8 2009-05 0.10000000
69 9 2009-05 0.10000000
70 10 2009-05 0.08000000
71 11 2009-05 0.06000000
72 12 2009-05 NA
73 13 2009-05 NA
74 14 2009-05 NA
75 15 2009-05 NA
76 1 2009-06 1.00000000
77 2 2009-06 0.46875000
78 3 2009-06 0.28125000
79 4 2009-06 0.18750000
80 5 2009-06 0.21875000
81 6 2009-06 0.15625000
82 7 2009-06 0.09375000
83 8 2009-06 0.09375000
84 9 2009-06 0.31250000
85 10 2009-06 0.09375000
86 11 2009-06 NA
87 12 2009-06 NA
88 13 2009-06 NA
89 14 2009-06 NA
90 15 2009-06 NA
91 1 2009-07 1.00000000
92 2 2009-07 0.46000000
93 3 2009-07 0.26000000
94 4 2009-07 0.20000000
95 5 2009-07 0.22000000
96 6 2009-07 0.20000000
97 7 2009-07 0.22000000
98 8 2009-07 0.14000000
99 9 2009-07 0.14000000
100 10 2009-07 NA
101 11 2009-07 NA
102 12 2009-07 NA
103 13 2009-07 NA
104 14 2009-07 NA
105 15 2009-07 NA
106 1 2009-08 1.00000000
107 2 2009-08 0.35483871
108 3 2009-08 0.29032258
109 4 2009-08 0.22580645
110 5 2009-08 0.19354839
111 6 2009-08 0.25806452
112 7 2009-08 0.12903226
113 8 2009-08 0.12903226
114 9 2009-08 NA
115 10 2009-08 NA
116 11 2009-08 NA
117 12 2009-08 NA
118 13 2009-08 NA
119 14 2009-08 NA
120 15 2009-08 NA
121 1 2009-09 1.00000000
122 2 2009-09 0.40540541
123 3 2009-09 0.37837838
124 4 2009-09 0.21621622
125 5 2009-09 0.35135135
126 6 2009-09 0.24324324
127 7 2009-09 0.21621622
128 8 2009-09 NA
129 9 2009-09 NA
130 10 2009-09 NA
131 11 2009-09 NA
132 12 2009-09 NA
133 13 2009-09 NA
134 14 2009-09 NA
135 15 2009-09 NA
136 1 2009-10 1.00000000
137 2 2009-10 0.31481481
138 3 2009-10 0.22222222
139 4 2009-10 0.24074074
140 5 2009-10 0.24074074
141 6 2009-10 0.12962963
142 7 2009-10 NA
143 8 2009-10 NA
144 9 2009-10 NA
145 10 2009-10 NA
146 11 2009-10 NA
147 12 2009-10 NA
148 13 2009-10 NA
149 14 2009-10 NA
150 15 2009-10 NA
151 1 2009-11 1.00000000
152 2 2009-11 0.24615385
153 3 2009-11 0.20000000
154 4 2009-11 0.22307692
155 5 2009-11 0.10000000
156 6 2009-11 NA
157 7 2009-11 NA
158 8 2009-11 NA
159 9 2009-11 NA
160 10 2009-11 NA
161 11 2009-11 NA
162 12 2009-11 NA
163 13 2009-11 NA
164 14 2009-11 NA
165 15 2009-11 NA
166 1 2009-12 1.00000000
167 2 2009-12 0.26153846
168 3 2009-12 0.27692308
169 4 2009-12 0.10769231
170 5 2009-12 NA
171 6 2009-12 NA
172 7 2009-12 NA
173 8 2009-12 NA
174 9 2009-12 NA
175 10 2009-12 NA
176 11 2009-12 NA
177 12 2009-12 NA
178 13 2009-12 NA
179 14 2009-12 NA
180 15 2009-12 NA
181 1 2010-01 1.00000000
182 2 2010-01 0.52631579
183 3 2010-01 0.27368421
184 4 2010-01 NA
185 5 2010-01 NA
186 6 2010-01 NA
187 7 2010-01 NA
188 8 2010-01 NA
189 9 2010-01 NA
190 10 2010-01 NA
191 11 2010-01 NA
192 12 2010-01 NA
193 13 2010-01 NA
194 14 2010-01 NA
195 15 2010-01 NA
196 1 2010-02 1.00000000
197 2 2010-02 0.19000000
198 3 2010-02 NA
199 4 2010-02 NA
200 5 2010-02 NA
201 6 2010-02 NA
202 7 2010-02 NA
203 8 2010-02 NA
204 9 2010-02 NA
205 10 2010-02 NA
206 11 2010-02 NA
207 12 2010-02 NA
208 13 2010-02 NA
209 14 2010-02 NA
210 15 2010-02 NA
211 1 2010-03 1.00000000
212 2 2010-03 NA
213 3 2010-03 NA
214 4 2010-03 NA
215 5 2010-03 NA
216 6 2010-03 NA
217 7 2010-03 NA
218 8 2010-03 NA
219 9 2010-03 NA
220 10 2010-03 NA
221 11 2010-03 NA
222 12 2010-03 NA
223 13 2010-03 NA
224 14 2010-03 NA
225 15 2010-03 NA")
The plot code using ggplot:
xlabels <- 1:15
ggplot(melted, aes(x=CohortPeriod, y=variable)) +
ggtitle('Retention by cohort') +
theme_bw() +
xlab('CohortPeriod') +
ylab('% of Cohort Purchasing') +
geom_tile(aes(fill = value), color='white') + geom_text(aes(label = round(value, 2)), size=2) +
scale_x_continuous(breaks = xlabels) +
scale_fill_gradient(low = 'white', high = '#2d1e3e', space = 'Lab', na.value = 'white') +
theme(axis.text.x=element_text(angle=0),
axis.ticks=element_blank(),
axis.line=element_blank(),
panel.border=element_blank(),
panel.grid.major=element_line(color='#eeeeee'))
If I do ggplot(melted, aes(x=CohortPeriod, y=rev(variable))) + ... it does flip the triangle but the y labels don't change with it.
How do I flip the triangle to ◸ while also sorting the y-axis labels correctly?
Upvotes: 0
Views: 5004
Answers (1)
Jack Brookes
Reputation: 3830
One way is to simply reorder the levels of the factor used for the variable column.
library(ggplot2)
# reverse factor levels
melted$variable <- factor(melted$variable, levels=rev(levels(melted$variable)))
xlabels <- 1:15
ggplot(melted, aes(x=CohortPeriod, y=variable)) +
ggtitle('Retention by cohort') +
theme_bw() +
xlab('CohortPeriod') +
ylab('% of Cohort Purchasing') +
geom_tile(aes(fill = value), color='white') + geom_text(aes(label = round(value, 2)), size=2) +
scale_x_continuous(breaks = xlabels) +
scale_fill_gradient(low = 'white', high = '#2d1e3e', space = 'Lab', na.value = 'white') +
theme(axis.text.x=element_text(angle=0),
axis.ticks=element_blank(),
axis.line=element_blank(),
panel.border=element_blank(),
panel.grid.major=element_line(color='#eeeeee'))
Upvotes: 4
Related Questions
- Heatmap: order y axis based on single x axis variable
- How to arrange a sloppy heatmap r
- reorder x-axis with heatmap in R
- Reorder axis of heatmap based on non-numeric third variable - ggplot2
- Heatmap rearrange columns
- ggplot2 heatmap : how to preserve the label order?
- Ordering heatmap rows by labRow
- Creating a second ggplot heatmap with the same row order as the first heatmap
- R - heatmap.2: reorder rows and columns
- How to order a heatmap in R by x or y?