-3

I'm new to R, but try to analyze dataset Here is the original link https://cache-default03g.cdn.yandex.net/download.yandex.ru/company/jobs/test_data_dreams.txt

My code is (I use R Studio 0.99.903 & R 3.3.1)

# get the data from url url <- "https://cache- default03g.cdn.yandex.net/download.yandex.ru/company/jobs/test_data_dreams.txt" testdata <-read.table(url, header = T, sep="\t") #install packages for text mining to analyze the queries install.packages("slam") install.packages("tm") library(tm) #convert unix to GMT testdata$timestamp..unix. <- as.POSIXct(as.numeric(as.character(testdata$timestamp..unix.)),origin="1970-01-01",tz="GMT") #delete some words testdata$query <- gsub("к чему снится ", "\\1", testdata$query) testdata$query <- gsub("к чему сниться ", "\\1", testdata$query) testdata$query <- gsub(" к чему снится", "\\1", testdata$query) testdata$query <- gsub(" к чему сниться", "\\1", testdata$query) testdata$query <- gsub("снится ", "\\1", testdata$query) testdata$query <- gsub(" к чему", "\\1", testdata$query)'

Now my data frame looks this way.

> head(testdata) timestamp..unix. query city 1 2016-02-04 10:15:13 волна вынесла на берег Москва 2 2016-02-24 10:28:53 бегать наперегонки Екатеринбург 3 2016-02-07 15:31:51 свадьба мужчине со своей женой Владикавказ 4 2016-02-05 08:06:24 иголка медицинская Тамбов 5 2016-02-16 15:21:16 давняя знакомая Калининград 6 2016-02-27 03:38:46 белый маленький котенок Новосибирск

Now I'm trying to plot queries to see their distribution during the daytime (also during the month) in general and for each city I have.

Could you please help me with the tool I should pick to read days and hours separately and plot not the query itself, but just the distribution of queries.

Thanks!

Yevgenyia Yusova
  • 21
  • 3
  • Hi @Yevgenyia, welcome to stack overflow. In order to get the help you need, the question you post should contain an example of the data set you are using (or simply a part of it, if it is too big), the attempt you made to solve the problem and the code you used. Have a look at the following link on [how to make a great R reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example). If you edit your question, I am sure we can help you out :) – thepule Aug 23 '16 at 15:55
  • @thepule Thank you for the link! It made everything more understandable ) – Yevgenyia Yusova Aug 24 '16 at 07:52

1 Answers1

0

You can use lubridate package, it is quite easy to extract days and hours from your dates and then to make a test on them. For example :

 # Add a column for day
df$day <- day(df$timestampcol)

# Add a column for the hour
df$hour <- hour(df$timestampcol)

For the plot I recommend you ggplot2 package, and here you will find a example plotting time series.

Community
  • 1
  • 1
Erick Díaz
  • 98
  • 2
  • 15