专栏名称: EasyCharts

EasyCharts，易图表，我们将定期推送各种数据可视化与分析教程，包括Excel（Power BI）、Origin、Sigmaplot、GraphPad、R、Python、Matlab、Tableau、D3.js等。

ggmatrix：R语言新型热力图包

EasyCharts · 公众号 · 前端 · 2020-01-03 00:01

正文

厚缊 业余的R语言可视化重度患者




    
个人博客:houyun.xyz邮箱：[email protected]转载本文（包括长期转载账号）必须联系厚缊授权

这个包的主要基于两个原因：昨天朋友告诉我被翻牌子把rectriangle整合进ggcor，我考虑了之后觉得放在ggcor不是很合适，就新开了一个包；另外一个是在做ggcor的过程中有很多新想法，需要先找个地方测试。当然，就目前来说，ggmatrix除了细节做的不够精致以外，其它的俨然有成为ggcor超集的趋势，看来我要一边捡芝麻一边仍西瓜了。庆幸的是，ggcor现在正在做互作网络一块的内容，也算两个有所区别，不重复劳动。 ggmatrix差不多就是一天的时间整完的，不完美是必然。

安装

# install.packages("devtools")devtools::install_github("houyunhuang/ggmatrix")

案例

先看被翻牌子的那个图吧。

library(ggmatrix)library(RColorBrewer)df                 y = rep(1:14, each = 11),                group = sample(LETTERS[1:5], 154, replace = TRUE),                values1 = rnorm(154, mean = 10, sd = 5),                values2 = rnorm(154, mean = -10, sd = 2),                stringsAsFactors = FALSE)ggplot(df, aes(x = x, y = y)) + geom_triangle()

ggplot(df, aes(x = x, y = y, fill = values1)) + geom_triangle()

ggplot(df, aes(x = x, y = y, fill = group)) + geom_triangle()

ggplot(df, aes(x = x, y = y, fill.upper = values1)) + geom_triangle() + scale_fill_upper_gradientn(colours = c("red", "white", "blue"))

ggplot(df, aes(x = x, y = y, fill.lower = values2)) + geom_triangle() + scale_fill_lower_gradientn(colours = c("#E9A3C9", "#F7F7F7", "#A1D76A"))

ggplot(df, aes(x = x, y = y, fill.upper = values1, fill.lower = values2)) + geom_triangle() + scale_fill_upper_gradientn(colours = c("red", "white", "blue")) + scale_fill_lower_gradientn(colours = c("#E9A3C9", "#F7F7F7", "#A1D76A"))

ggplot(df, aes(x = x, y = y, fill.upper = values1, fill.lower = values2)) + geom_triangle(mode = "lb-rt") + scale_fill_upper_gradientn(colours = c("red", "white", "blue")) + scale_fill_lower_gradientn(colours = c("#E9A3C9", "#F7F7F7", "#A1D76A"))

ggplot(df, aes(x = x, y = y, fill.upper = values1, fill.lower = values2)) + geom_triangle() + scale_fill_upper_gradientn(colours = brewer.pal(5, "Greys")) + scale_fill_lower_gradientn(colours = brewer.pal(5, "YlGnBu"))

ggplot(df, aes(x = x, y = y, fill.upper = values1, fill.lower = values2)) + geom_triangle(r0 = 0.4) + scale_fill_upper_gradientn(colours = c("red", "white", "blue")) + scale_fill_lower_gradientn(colours = c("#E9A3C9", "#F7F7F7", "#A1D76A"))

ggplot(df, aes(x = x, y = y, fill.upper = group, fill.lower = values2)) + geom_triangle() + scale_fill_upper_manual(values = brewer.pal(5, "Set1")) + scale_fill_lower_gradientn(colours = c("#E9A3C9", "#F7F7F7", "#A1D76A"),                            guide = guide_legend())

矩阵热图

之所以叫ggmatrix，主要是突出这个在做矩阵热图上的一些小优势，而产生优势的核心是matrixs_to_df()函数。

matrixs_to_df(mtcars = mtcars)## # A tibble: 352 x 5##    row.names         col.names mtcars row.id col.id##  *                        ##  1 Mazda RX4         mpg         21       32      1##  2 Mazda RX4 Wag     mpg         21       31      1##  3 Datsun 710        mpg         22.8     30      1##  4 Hornet 4 Drive    mpg         21.4     29      1##  5 Hornet Sportabout mpg         18.7     28      1##  6 Valiant           mpg         18.1     27      1##  7 Duster 360        mpg         14.3     26      1##  8 Merc 240D         mpg         24.4     25      1##  9 Merc 230          mpg         22.8     24      1## 10 Merc 280          mpg         19.2     23      1## # … with 342 more rowsm1 m2 m3 matrixs_to_df(m1 = m1, m2 = m2, m3 = m3)## # A tibble: 400 x 7##    row.names col.names     m1     m2    m3 row.id col.id##  *                  ##  1 row1      col1      -0.639  -9.31  13.0     20      1##  2 row2      col1       0.494 -10.4   12.4     19      1##  3 row3      col1      -0.296  -9.88  11.1     18      1##  4 row4      col1       1.40   -9.89  12.3     17      1##  5 row5      col1       1.37   -9.12  11.1     16      1##  6 row6      col1       0.122  -8.87  13.7     15      1##  7 row7      col1      -0.525  -9.63  11.5     14      1##  8 row8      col1      -1.95  -11.2   13.0     13      1##  9 row9      col1      -2.56  -10.1   12.0     12      1## 10 row10     col1      -0.169 -12.3   11.3     11      1## # … with 390 more rows

支持按照输入的第一个矩阵聚类。

matrixs_to_df(m1 = m1, m2 = m2, m3 = m3, cluster = TRUE)## # A tibble: 400 x 7##    row.names col.names     m1     m2    m3 row.id col.id##  *                  ##  1 row18     col7      -1.36  -10.0   12.9     20      1##  2 row15     col7      -0.641  -8.22  12.8     19      1##  3 row5      col7      -0.227 -10.4   12.7     18      1##  4 row11     col7      -1.14   -9.68  11.8     17      1##  5 row13     col7       0.354 -10.4   12.3     16      1##  6 row2      col7      -0.162  -9.21  11.2     15      1##  7 row10     col7       0.430 -10.5   12.4     14      1##  8 row6      col7      -0.323  -8.08  10.9     13      1##  9 row16     col7       2.07  -10.5   12.9     12      1## 10 row9      col7      -1.54  -10.3   11.4     11      1## # … with 390 more rowsmatrixs_to_df(m1 = m1, m2 = m2, m3 = m3, cluster = TRUE, cluster.type = "row")## # A tibble: 400 x 7##    row.names col.names      m1     m2    m3 row.id col.id##  *                    ##  1 row18     col1      -0.467  -10.6   13.3     20      1##  2 row15     col1      -0.0183  -9.14  10.6     19      1##  3 row5      col1       1.37    -9.12  11.1     18      1##  4 row11     col1      -0.0757 -11.8   13.3     17      1##  5 row13     col1       1.93    -9.31  12.0     16      1##  6 row2      col1       0.494  -10.4   12.4     15      1##  7 row10     col1      -0.169  -12.3   11.3     14      1##  8 row6      col1       0.122   -8.87  13.7     13      1##  9 row16     col1       1.15   -10.1   11.3     12      1## 10 row9      col1      -2.56   -10.1   12.0     11      1## # … with 390 more rowsmatrixs_to_df(m1 = m1, m2 = m2, m3 = m3, cluster = TRUE, cluster.type = "col")## # A tibble: 400 x 7##    row.names col.names     m1     m2    m3 row.id col.id##  *                  ##  1 row1      col7       0.779  -9.04  12.6     20      1##  2 row2      col7      -0.162  -9.21  11.2     19      1##  3 row3      col7      -1.48   -9.76  12.6     18      1##  4 row4      col7      -1.05   -9.36  12.1     17      1##  5 row5      col7      -0.227 -10.4   12.7     16      1##  6 row6      col7      -0.323  -8.08  10.9     15      1##  7 row7      col7       0.795  -9.08  12.2     14      1##  8 row8      col7      -0.655 -10.6   10.5     13      1##  9 row9      col7      -1.54  -10.3   11.4     12      1## 10 row10     col7       0.430 -10.5   12.4     11      1## # … with 390 more rows

在传入矩阵时，一定要用name1 = mat1的形式，不然数据会丢失。

matrixs_to_df(mtcars)## # A tibble: 352 x 4##    row.names         col.names row.id col.id##  *                      ##  1 Mazda RX4         mpg           32      1##  2 Mazda RX4 Wag     mpg           31      1##  3 Datsun 710        mpg           30      1##  4 Hornet 4 Drive    mpg           29      1##  5 Hornet Sportabout mpg           28      1##  6 Valiant           mpg           27      1##  7 Duster 360        mpg           26      1##  8 Merc 240D         mpg           25      1##  9 Merc 230          mpg           24      1## 10 Merc 280          mpg           23      1## # … with 342 more rows

矩阵热图可视化

初始化函数是ggmat，可以直接把要可视化的矩阵丢进去，底层调用matrixs_to_df()函数转换成数据框。

library(ggcor)ggmat(mtcars = mtcars) + geom_triangle(aes(fill = mtcars))

ggmat(m1 = m1) + geom_square(aes(fill = m1, r0 = m1)) + scale_fill_gradient2n(colours = RColorBrewer::brewer.pal(11, "RdBu"),                       midpoint = mean(m1)) + scale_radius_area(midpoint = mean(m1))

ggmat(m1 = m1) + geom_ellipse2(aes(fill = m1, r0 = m1)) + scale_fill_gradient2n(colours = RColorBrewer::brewer.pal(11, "RdBu"),                       midpoint = mean(m1)) + scale_radius_area(midpoint = mean(m1))

      
         如需联系EasyCharts团队
         
         请加微信：
        
          EasyCharts

      
              《R语言数据可视化之美》

ggmatrix：R语言新型热力图包

正文

请到「今天看啥」查看全文