r - r：带有自定义图例和永久显示国名的 Choropleth

Question

以下代码根据层次聚类创建 Choropleth 地图（使用hclust()和cutree()）：

library(plotly)
library(cluster)

hc <- hclust(dist(df), method = "complete")
df$member <- cutree(hc, 5)

cluster.means = (as.data.frame(aggregate(df[,-1], list(cluster= df$member), mean)))[,-4]

g = list(
  scope = 'africa',
  showframe = T,
  showland = T,
  landcolor = toRGB("white")
)

plot_ly(df, z = member, type = 'choropleth', locations = Country,
              locationmode = 'country names', text = Country, hoverinfo = "text") %>% 
        layout(geo = g, title = "Energy markets in Africa")

现在我想改变两件事：

1. 永久显示国家名称，即不仅在 RStudio 中的交互式图表上悬停时。我试图将这篇文章的答案应用于我的，但没有成功。
2. 一个不连续的尺度。理想情况下，我希望一个集群完全没有比例，但每个集群有 5 个框，每个框具有以下三个特征：（1）框的颜色应根据其在等值线中的相应颜色（2）每个框应包含数字分配给它的国家，即相应的集群 (3) 每个框应包含两个变量 (X1, X2) 的各自集群均值cluster.means

我附上了这样一个我所想的示例图（它不必看起来完全相同 - 只是为了传达这个想法）。

任何帮助、建议、tipp 都非常感谢！

（缩放）数据如下所示：

df <- structure(list(Country = structure(1:50, .Label = c("Angola", 
"Benin", "Botswana", "Burkina Faso", "Burundi", "Cabo Verde", 
"Cameroon", "Central African Republic", "Chad", "Comoros", "Congo", 
"Cote d'Ivoire", "Democratic Republic of Congo", "Djibouti", 
"Equatorial Guinea", "Eritrea", "Ethiopia", "Gabon", "Gambia", 
"Ghana", "Guinea", "Guinea-Bissau", "Kenya", "Lesotho", "Liberia", 
"Madagascar", "Malawi", "Mali", "Mauritania", "Mauritius", "Mozambique", 
"Namibia", "Niger", "Nigeria", "Reunion", "Rwanda", "Sao Tome and Principe", 
"Senegal", "Seychelles", "Sierra Leone", "Somalia", "South Africa", 
"South Sudan", "Sudan", "Swaziland", "Tanzania", "Togo", "Uganda", 
"Zambia", "Zimbabwe"), class = "factor"), X1 = c(-0.18, -1.03, 
0.6, 1.55, 0.22, 0.26, 0.76, 2.15, -1.43, 0.99, 1.79, -0.39, 
1.73, 1.57, 1.11, -0.09, -1.49, -0.46, -0.48, -1.22, -0.78, -1.46, 
-1.22, 0.35, 0.45, 1.29, -1.37, -0.61, 0.92, -1.3, 0.42, -1.18, 
1.4, -0.83, 0.06, -0.76, -0.19, -0.37, -0.63, 0.64, 0.93, 0.33, 
-0.76, -0.21, -0.59, -0.41, -0.74, 0.39, -1.1, 1.35), X2 = c(-0.22, 
-0.42, 0.72, -0.59, -1.27, 0.64, -1.35, -1.4, -0.35, -1.43, 1.07, 
-0.01, -0.51, 0.11, 1.14, -0.89, 0.77, 1.45, -1.67, -0.83, 0.71, 
0.92, 1.63, 1.68, 0.23, -0.18, 0.07, 0.8, -0.02, 0.82, -0.72, 
-0.41, -0.26, 0.02, -1.68, 1.67, 0.18, 0.98, 1.45, 0.31, -1.23, 
-1.38, -0.63, 1.41, -0.12, 0, -1.3, -1.64, 0.21, 1.52)), .Names = c("Country", 
"X1", "X2"), row.names = c(NA, -50L), class = "data.frame")

Answer 1

“理想情况下，我希望完全没有比例”：这是通过将参数添加showscale=F到plot_ly().

“国家名称的永久显示”：这是通过添加一个scattergeo绘图层来完成的，如本文所述，略有不同。因为您的数据框包含一列国家名称，但不是国家代码之一，您需要添加参数locationmode = 'country names'

p <- plot_ly(df, z = member, type = 'choropleth', 
             locations = Country, locationmode = 'country names', 
             text = Country, hoverinfo = "text", 
             showscale=F, inherit =F) %>% 
     layout(geo = g, title = "Energy markets in Africa") %>%
     add_trace(type="scattergeo",
             locationmode = 'country names', locations = Country, 
             text = Country, mode="text", 
             textfont = list(color=rgb(1,0.5,0.3),  size =12)) 

您将需要尝试各种配色方案，以找到一种能有效显示集群的配色方案，同时使永久国家名称标签相对可见。当地图缩小时，永久国家名称标签将相互重叠，您可能无能为力。通过额外的努力，您可以为每个标签指定一个经纬度位置，而不是依赖于它们的默认位置，这可能有助于通过更好地分离标签来改善地图的外观。不过，我认为，由于您要求提供永久标签，因此您打算仅以合适的缩放比例使用地图，使它们看起来不错。

要添加您描述的文本框，您可以使用注释布局（每个集群一个注释）。我在这里展示了如何添加一个带有紫色边框和第二个绿松石色的边框。您需要将其调整为最终使用的配色方案。

box1 <- list(
  x = 0.3,
  y = 0.5,
  yanchor = "top",
  borderpad = 2,
  bordercolor = rgb(0.5,0.1,0.5),     # set this same as color of cluster 1
  borderwidth = 5,
  text = paste("1. Cluster,  ", 
               sum(df$member==1),     # number of countries in cluster
               " Countries<br>X1 = ", # use <br> for line breaks
               format(round(cluster.means[1,]$X1, 2), nsmall = 2), 
               "<br>X2 = ", 
               format(round(cluster.means[1,]$X2, 2), nsmall = 2)),
  align = "left",
  showarrow = F)

box2 <- list(
  x = 0.3,
  y = 0.4,
  yanchor = "top",
  borderpad = 2,
  bordercolor = rgb(0.1,0.5,0.5),
  borderwidth = 5,
  text = paste("2. Cluster,  ", 
               sum(df$member==2), 
               " Countries<br>X1 = ", 
               format(round(cluster.means[2,]$X1, 2), nsmall = 2), 
               "<br>X2 = ", 
               format(round(cluster.means[2,]$X2, 2), nsmall = 2)),
  align = "left",
  showarrow = F)

p %>% layout(annotations = list(box1, box2))