Graphing My Daily Phone Use

January 27, 2019
By

(This article was first published on Category R on Roel's R-tefacts, and kindly contributed to R-bloggers)

How many times do I look at my phone? I set up a small program on my phone
to count the screen activations and logged to a file. In this post I show
what went wrong and how to plot the results.

The data

I set up a small program on my phone that counts every day how many times I use
my phone (to be specific, it counts the times the screen has been activated).

My data looks like this:

"screen_log";"1-19-19";"17.30";"7"
"screen_log";"1-19-19";"17.33";"8"
"screen_log";"1-19-19";"17.36";"9"
"screen_log";"1-19-19";"17.38";"10"

To account for comma use and possible problems I set up the program on my
phone to write a “;”-seperated file that records screen_log, the date, the
time and the current value of screen_count. Every day around 12 o clock it reset
the counter to 0.
I thought it would be cool to compare different days.

The problems

I started the data collection on januari 19th around 17:00h, so the first day
is already halfway through.
For reasons I cannot fathom, sometimes the system date is recorded in the USA
style MONTH-DAY-YEAR and sometimes in the rest-of-the-world style of DAY-MONTH-YEAR.
I wish I could set it to YEAR-MONTH-DAY (ISO 8601).

Reading in the data

I use read_csv2, which expects “;” as a seperator and never converts text to factor.
This particular textfile has no headers, so I tell R what to call the columns.

library(tidyverse) # what else
screenlog <- read_csv2("data/screenlog.csv",col_names = c("type","date","time","counter"))

Data cleaning

I have to deal with the different time formats, so I set up a regex that works
only with Januari, if it detects -01-19 it pulls out the numbers before that,
if it detects the other variant it takes the second part.
I combine the date and time into a timestamp and pull out the hours and minutes,
before combining the hours and minutes into HMS time class.
Finally I remove anything over 23 hours, because in that period the counter is
reset.

screenlog <-   
  screenlog %>% 
    mutate(
        day = case_when( 
            str_detect(date, "[0-9]{1,2}-01-19") ~ 
                str_replace(date, "([0-9]{1,2})-01-19","\\1"),
            str_detect(date, "1-[0-9]{1,2}-19") ~ 
              str_replace(date, "1-([0-9]{1,2})-19", "\\1") ,
            TRUE ~  NA_character_
            ),
        timestamp = paste0("2019-01-",day, " ",time),
        timestamp = as.POSIXct(timestamp,tz = "GMT", format = "%Y-%m-%d %H.%M"),
        hours = str_replace(time,"\\.[0-9]{1,2}", "") %>% as.numeric(),
        minutes = str_replace(time,"[0-9]{1,2}\\.", "") %>% as.numeric(),
        time = hms::hms(hours = hours, minutes = minutes)
        ) %>% 
    filter(hours < 23)

How does it look?

First an overview:

screenlog %>% 
    ggplot(aes(timestamp, counter, color = day))+
    geom_step()+
    ggtitle("Times I looked at my screen during vacation", subtitle = "daily values")+
    theme_light()
Daily cumulative screen looking values

Daily cumulative screen looking values

But that is difficult to compare, so I also show them overlayed:

screenlog %>% 
    ggplot(aes(time, counter, group = day, color = day))+
    geom_step()+
    ggtitle("Times I looked at my screen during vacation", subtitle = "overlay plot")+
    theme_light()
overlay of cumulative screen lookings every day on the same hourly scale

overlay of cumulative screen lookings every day on the same hourly scale

Fin

The only remaining question is: what did I do on the 25th that I looked soooo (326 times) many
times on my screen?
Is there a bug in the counting? Was I really bored, did I take a lot of photo’s?
I was in the Botanical Gardens of Malaga and did take a lot of pictures with my
phone.

To leave a comment for the author, please follow the link and comment on their blog: Category R on Roel's R-tefacts.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.

Search R-bloggers


Sponsors

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)