同一图表上的多条频率线，其中 y 是字符值

Question

我正在尝试创建图表类型按年份出现次数的频率图。 我已经使用 ggplot2 一段时间了，但我认为这超出了我的能力范围（我刚刚开始使用 R）

我附上了我希望结果是什么样子的示意图。我遇到的其他问题之一是，图形类型很多年都没有出现。如果该图表类型当年没有出现，是否有办法排除该图表类型？

例如，在 1940 年，没有“社会图”，我不想在 0 处有一堆线...

year <- c("1940","1940","1940","1940","1940","1940","1940","1940","1940","1940","1940","1941","1941","1941","1941","1941","1941","1941","1941","1941","1941","1941","1941","1941","1941")
type <- c("Line","Column", "Stacked Column", "Scatter with line", "Scatter with line", "Scatter with line", "Scatter with line", "Map with distribution","Line","Line","Line","Bar","Bar","Stacked bar","Column","Column","Sociogram","Sociogram","Column","Column","Column","Line","Line","Line","Line")
ytmatrix <- cbind(as.Date(as.character(year), "%Y", type))

如果有些事情没有意义，请告诉我。 StackOverflow 正迅速成为我最喜欢的网站之一！

感谢， Jon

---

这是我到目前为止所拥有的...... 再次感谢您的所有帮助！

这就是我的做法（我还不能共享数据文件，因为我们希望将其用于出版物，但 ggplot 区域可能更有趣，尽管我并没有真正做任何新的事情/帖子中没有讨论）：

AJS = read.csv(data) #read in file
Type = AJS[,17] #select and name "Type" column from csv
Year = AJS[,13] #select and name "Year" column from csv
Year = substr(Year,9,12) #get rid of junk from year column
Year = as.Date(Year, "%Y") #convert the year character to a date
Year = format(Year, "%Y") #get rid of the dummy month and day
Type = as.data.frame(Type) #create data frame
yt <- cbind(Year,Type) #bind the year and type together
library(ggplot2) 

trial <- ggplot(yt, aes(Year,..count.., group= Type)) + #plot the data followed by aes(x-  axis, y-axis, group the lines)
geom_density(alpha = 0.25, aes(fill=Type)) +
opts(axis.text.x = theme_text(angle = 90, hjust = 0)) + #adjust the x axis ticks to horizontal
opts(title = expression("Trends in the Use of Visualizations in The American Journal of Sociology")) + #Add title
scale_y_continuous('Appearances (10 or more)') #change Y-axis label
trial

Answer 1

这可能是一个更有趣的数据框来进行实验：

df1 <- data.frame(date = as.Date(10*365*rbeta(100, .5, .1)),group="a")
 df2 <- data.frame(date = as.Date(10*365*rbeta(50, .1, .5)),group="b")
 df3 <- data.frame(date = as.Date(10*365*rbeta(25, 3,3)),group="c")
 dfrm <- rbind(df1,df2,df3)

我认为使用帮助（stat_密度）页面中的示例可以工作，但事实并非如此：

m <- ggplot(dfrm, aes(x=date), group=group)
m+ geom_histogram(aes(y=..density..)) + geom_density(fill=NA, colour="black")

但是，我在搜索 hte 档案中找到的一个示例发现了 @Hadley Wickham 的帖子确实有效：

m+ geom_density(aes(fill=group), colour="black")