How to Plot line chart using R for time-series analysis - r

Question

I am trying to plot a line chart using Date-time and no of tweets at that period of date and time in R.
library(ggplot2)
df1 <- structure(list(Date = structure(c(1L, 1L, 2L, 1L, 1L, 1L), .Label = c("2020-03-12",
"2020-03-13"), class = "factor"), Time = structure(c(1L, 1L, 2L,
3L, 4L, 5L), .Label = c("00:00:00Z", "00:00:01Z", "00:10:04Z",
"00:25:12Z", "01:00:02Z"), class = "factor"), Text = structure(c(5L,
3L, 6L, 4L, 2L, 1L), .Label = c("The images of demonstrations and gathering", "Premium policy get activate by company abc",
"Launches of rocket", "Premium policy get activate by company abc",
"Technology makes trend", "The images of demonstrations and gatherings",
"Weather forecasting by xyz"), class = "factor")), class = "data.frame", row.names = c(NA,
-6L))
ggplot(df1, aes(x = Date, y = text(count)) + geom_line(aes(color = variable), size = 1)
I tried the above code to plot desired result but got an error. Dataset given like that in csv format.
Date Time Text
2020-03-12 00:00:00Z The images of demonstrations and gatherings
2020-03-12 00:00:00Z Premium policy get activate by company abc
2020-03-12 00:00:01Z Weather forecasting by xyz
2020-03-12 00:10:04Z Technology makes trend
2020-03-12 00:25:12Z Launches of rocket
2020-03-12 01:00:02Z Government launch new policy to different sector improvement
I have a dataset of nearly 15 days and want to plot the line chart to visualize the number of tweets (given in text column) to see the trend of tweets on different time and date.

Answer 1

df1 <- structure(list(Date = structure(c(1L, 1L, 2L, 1L, 1L, 1L), .Label = c("3/12/2020",
"3/13/2020"), class = "factor"), Time = structure(c(1L, 1L, 2L,
3L, 4L, 5L), .Label = c("00:00:00Z", "00:00:01Z", "00:10:04Z",
"00:25:12Z", "01:00:02Z"), class = "factor"), Text = structure(c(5L,
3L, 6L, 4L, 2L, 1L), .Label = c("Government launch new policy to different sector",
"Launches of rocket", "Premium policy get activate by company abc",
"Technology makes trend", "The images of demonstrations and gatherings",
"Weather forecasting by xyz"), class = "factor"), X = structure(c(1L,
1L, 1L, 1L, 1L, 2L), .Label = c("", "improvement"), class = "factor")), class = "data.frame", row.names = c(NA,
-6L))
Creating the dataset df1 as above then running this gives you required plot for hour
library(tidyverse)
library(lubridate)
df1 %>%
mutate(Time=hms(Time),
Date=mdy(Date),
hour=hour(Time)) %>%
count(hour) %>%
ggplot(aes(hour,n,group=1))+geom_line()+geom_point()

Answer 2

Is this what you are after?
library(dplyr)
library(lubridate)
library(stringr)
library(ggplot2)
Answer with your data
To demonstrate data wrangling.
# your data;
df1 <- structure(list(Date = structure(c(1L, 1L, 2L, 1L, 1L, 1L),
.Label = c("2020-03-12","2020-03-13"),
class = "factor"),
Time = structure(c(1L, 1L, 2L,3L, 4L, 5L),
.Label = c("00:00:00Z", "00:00:01Z", "00:10:04Z","00:25:12Z", "01:00:02Z"),
class = "factor"),
Text = structure(c(5L,3L, 6L, 4L, 2L, 1L),
.Label = c("The images of demonstrations and gathering", "Premium policy get activate by company abc",
"Launches of rocket", "Premium policy get activate by company abc",
"Technology makes trend", "The images of demonstrations and gatherings", "Weather forecasting by xyz"), class = "factor")),
class = "data.frame", row.names = c(NA,-6L))
# data wrangle
df2 <-
df1 %>%
# change all variables from factors to character
mutate_all(as.character) %>%
mutate(Time = str_remove(Time, "Z$"), #remove the trailing 'Z' from Time values
dt = ymd_hms(paste(Date, Time, sep = " ")), # change text into datetime format using lubridtate::ymd_hms
dt = ceiling_date(dt, unit="hour")) %>% # round to the end of the named hour, separated for clarity
group_by(dt) %>%
summarise(nr_tweets = n())
# plot
p1 <- ggplot(df2, aes(dt, nr_tweets))+
geom_line()+
scale_x_datetime(date_breaks = "1 day", date_labels = "%d/%m")+
ggtitle("Data from question `df1`")
Answer with made up large dataset
tib <- tibble(dt = sample(seq(ISOdate(2020,05,01), ISOdate(2020,05,15), by = "sec"), 10000, replace = TRUE),
text = sample(c(letters[1:26], LETTERS[1:26]), 10000, replace = TRUE))
tib1 <-
tib %>%
mutate(dt = round_date(dt, unit="hour"))%>%
group_by(dt) %>%
summarise(nr_tweets = n())
p2 <- ggplot(tib1, aes(dt, nr_tweets))+
geom_line()+
scale_x_datetime(date_breaks = "1 day", date_labels = "%d/%m")+
ggtitle("Result using `tib` data made up to answer the question")
p1/p2
Created on 2020-05-13 by the reprex package (v0.3.0)