{r setup, include=FALSE} knitr::opts_chunk$set(eval = FALSE) ggplot中，geom_jitter()和position_jitter()都是处理点的处理图像重叠的问题。
如图，
{r message=FALSE, warning=FALSE} library(tidyverse) library(lubridate) data1 <- tibble::tribble( ~logtype, ~inserttime, "APP主登录", "2017-07-29 13:10:31", "延伸登录", "2017-08-29 19:03:12", "延伸登录", "2017-08-29 19:07:17", "独立登录", "2017-08-31 17:28:07", "独立登录", "2017-08-31 17:28:07", "独立登录", "2017-08-31 17:45:24", "独立登录", "2017-09-05 02:31:37", "独立登录", "2017-09-05 02:32:15", "独立登录", "2017-09-05 02:34:50", "M站主登录", "2017-09-05 02:52:02 ) %>% rename(item=logtype, start = inserttime) %>% mutate(start = ymd_hms(start)) data1 %>% # mutate(start = as.Date(start)) %>% ggplot(aes(x = start, y = fct_reorder(item, start, .desc = TRUE)) ) + geom_point( # position=position_jitter(width=0.5,height=0), size = 6, alpha = 0.5, col = 'gold' ) + xlab(NULL) + ylab(NULL) + theme_light() + theme( legend.position = "none", panel.border = element_blank(), ) + theme(text = element_text(family = 'STKaiti', size=10))
发现图中有图重合现象，这是因为独立登录发生时间太过于相近。
处理的方法是将比较相近的相同登录，转为同一时间，函数为 mutate(start = as.Date(start))
然后加入jitter函数， position=position_jitter(width=0.5,height=0)。 其中，height=0表示控制点不纵向扰动。
{r message=FALSE, warning=FALSE} data1 %>% mutate(start = as.Date(start)) %>% ggplot(aes(x = start, y = fct_reorder(item, start, .desc = TRUE)) ) + geom_point( position=position_jitter(width=0.5,height=0), size = 6, alpha = 0.5, col = 'gold' ) + xlab(NULL) + ylab(NULL) + theme_light() + theme( legend.position = "none", panel.border = element_blank(), ) + theme(text = element_text(family = 'STKaiti', size=10))