CODEWITHSUNDEEP
rlinechart
Asmita Poddar
Asmita Poddar

Reputation: 532

Creating a line chart in R across time

I have a data frame like the following:

     user_id                         track_id          created_at
1   81496937 cd52b3e5b51da29e5893dba82a418a4b 2014-01-01 05:54:21    
2 2205686924 da3110a77b724072b08f231c9d6f7534 2014-01-01 05:54:22
3  132588395 ba84d88c10fb0e42d4754a27ead10546 2014-01-01 05:54:22
4   97675221 33f95122281f76e7134f9cbea3be980f 2014-01-02 05:54:24
5   17945688 b5c42e81e15cd54b9b0ee34711dedf05 2014-01-02 05:54:24
6  452285741 8bd5206b84c968eda0af8bc86d6ab1d1 2014-01-02 05:54:25

I want to create a line chart in R showing the number of user_id across days. I want to know how many user_id are present per day and create a plot of that. How do I do it?

Upvotes: 1

Views: 231

Answers (1)

Shuiping Chen
Shuiping Chen

Reputation: 426

First of all, you should know how to process date and time in R. I strongly recommend the lubridate package.

library(lubridate)
t <- ymd_hms("20170621111800")
dt <- floor_date(t, unit='day')
dt

Then you need to learn how to manipulate a data frame in R. I usually use dplyr package because it is quite simple to learn and the code is easy to read.

library(dplyr)
new_df <- df %>% 
  mutate(dt=floor_date(ymd_hms(created_at, unit='day'))) %>%
  group_by(dt) %>%
  summarise(user_cnt=n_distinct(user_id))
new_df

At last, you need to learn how to plot a data frame in R. I personally prefer to use ggplot2 to do this task.

library(ggplot2)
p <- ggplot(new_df) + geom_line(aes(x=dt, y=user_cnt))
p

Now you will see a picture showed in the bottom right panel if you use RStudio to run the code. Furthermore, you could use plotly package to change the static image to a dynamic chart!

library(plotly)
ggplotly(p)

Upvotes: 3

Related Questions