R

Question

我有一个带有几个二进制变量的表。是否可以制作一张表格，使所有带有1的单元格都为黑色。像这样

Answer 1

使用gridExtra创建一个gtable。

library(grid)
library(gridExtra)

set.seed(1)

sample.data <- matrix(data=sample(0:1, 5*5, r=TRUE), nrow=5, ncol=5)
colnames(sample.data) <- LETTERS[1:5]
rownames(sample.data) <- letters[1:5]


# for cut() to work data has to be of class matrix or table.
# data.frame will cause error    
fcol <- as.character(cut(
          sample.data, c(0, 0.5, 1), include.lowest=TRUE, 
          labels=c("black", "white")))

bcol <- as.character(cut(
          sample.data, c(0, 0.5, 1), include.lowest=TRUE, 
          labels=c("white", "black")))

# building a theme.
# core specifies the features of the main field,
# rowhead the row titles, and colhead the column titles.
# fg_params specifies the features of the foreground, namely the text,
# bg_params specifies the features of the background, like the fill
# and border colors.
t1 <- ttheme_minimal(base_size=20, 
        core=list(
          fg_params=list(col=fcol),
          bg_params=list(fill=bcol, col="black", lwd=2)
        ),
        rowhead=list(
          fg_params=list(fontface=1)
        ),
        colhead=list(
          fg_params=list(fontface=1, rot=90)
        )
      )

grid.newpage()
grid.table(sample.data, theme=t1)

Answer 2

我们将使用图像中的数据（尽管这可能不是您的用例）。这将模拟它：

library(tidyverse)

set.seed(2018-11-20)

# expand.grid just creates a data frame out of all the combinations
# of the provided vectors.

# tbl_df just turns the output into a data frame that (oversimplifying)
# prints better

# set_names is just setting column names

# the mutate line is where we generate the 0/1s

expand.grid(
  sprintf("Address%s", 1:9),
  sprintf("Transaction%s", 1:16),
  stringsAsFactors = FALSE
) %>%
  tbl_df() %>%
  set_names(c("address", "transaction")) %>%
  mutate(value = sample(0:1, n(), replace=TRUE)) -> xdf

xdf
## # A tibble: 144 x 3
##    address  transaction  value
##    <chr>    <chr>        <int>
##  1 Address1 Transaction1     0
##  2 Address2 Transaction1     1
##  3 Address3 Transaction1     1
##  4 Address4 Transaction1     0
##  5 Address5 Transaction1     1
##  6 Address6 Transaction1     1
##  7 Address7 Transaction1     0
##  8 Address8 Transaction1     0
##  9 Address9 Transaction1     1
## 10 Address1 Transaction2     0
## # ... with 134 more rows

如果您的数据是“宽”还是“长”（如上所述），请查看tidyr::gather函数。

现在，这只是更多的数据争执和一些基本的ggplot2。首先，我们将使用您的逻辑并为每个单元格定义单元格填充和文本颜色。并且，由于我们希望以相反的方向绘制X轴，因此我们也对其进行了排序，以便ggplot2可以满足这种需求。

mutate(
  xdf,
  fill = ifelse(value == 1, "black", "white"),
  color = ifelse(value == 1, "white", "black"),
  address = factor(address, levels = sort(unique(address), decreasing = TRUE))
) -> xdf

现在，我们使用平铺层，并在其上面放置一个文本层。我们在上面的mutate调用中使用了“真实的”颜色名称，因此我们可以通过用I()包围映射的列美感来利用ggplot2的“ hack”，这告诉ggplot2按原样使用值，并且也不必费心给我们一个图例（如果您愿意，我们可以生成一个图例，但这不在稀疏问题中）。

然后，我们删除标签和地块之间的一些空白，将Y标签粘贴在顶部并旋转它们（尽管这不是一个好习惯，因为让人们扭动头来阅读您的地块并不酷）。

ggplot(xdf, aes(x = transaction, y = address)) +
  geom_tile(
    aes(fill = I(fill)),
    color = "#2b2b2b", size=0.125,
  ) +
  geom_text(
    aes(label = value, color = I(color))
  ) +
  scale_x_discrete(expand=c(0,0), position = "top") +
  scale_y_discrete(expand=c(0,0)) +
  labs(x = NULL, y = NULL) +
  hrbrthemes::theme_ipsum_rc(grid="XY") +
  theme(axis.text.x = element_text(angle=90, vjust=0.5, hjust=1))