Peaks in Internet traffic: Analysis of data demand through time-use diaries and Android device-use

This dataset concerns the timing of Internet traffic associated with
everyday activities. It provides evidence about patterns of data demand
from two different sources: time-use diaries and analysis of data usage by
Android smartphones and tablets. 1. Time-use diaries File:
Diaries_TVAndTimeOnline This is a summarised record produced from 16
time-use diaries completed by volunteers in the UK, for one week falling
sometime between July and October 2016 (not the same week). The file
summarises the time of day (averaged across the week) when three
activities were recorded: 'TV watching', 'online
activities’ conducted anywhere, and those ‘online activities’ conducted at
home. This was recorded by the diarist in two ways: 1) For two days of the
week, diarists recorded all their activities in their own words on a
schedule sheet. Categories were assigned for each half hour interval by
the analyst based on what diarists had written. 2) For the other five days
of the week, diarists recorded their activities by marking the relevant
times in a table for when they were online, watching TV, and at home. A
summary of the ages, working and living status of the participants is
given on the notes page. 2. Smartphone and tablet data usage File:
DeviceAnalyzer_SmartphoneTabletData This is a summary of the results of a
secondary analysis conducted, on a pre-existing data set generated by the
ordinary use of smartphones and tablets. The data was collected through an
Android application, Device Analyzer, available on the Google Play store
which captures usage statistics about the device in the background, as it
is used by volunteers who have installed the app. This includes volumes of
data used by particular applications. The application was designed by
researchers at Cambridge University who curate and share data with other
researchers, where volunteers have given their permission for this. More
information and instructions on how to access the Device Analyzer dataset
are provided here: https://deviceanalyzer.cl.cam.ac.uk/. We analysed a
subset of this data where: (1) logs consisted of at least 14 days with the
latest data collected on or after 1st January 2014 up until 21 September
2016, (2) had a network-based location in the UK or Ireland for at least
half of the contribution days, and (3) used apps or demanded data during
their logging period. This resulted in 398 devices. Due to anonymity
restrictions associated with this dataset, we do not know the age, gender
or device make-up of this sample; all those who participated indicate they
were over 18 years old. The file included here has been produced from
scripts written by one of the authors and run on the dataset in
co-ordination with researchers at Cambridge University. It presents the
hourly distribution of data demanded by the device (upload and download)
by day of the week and across app categories (as an average of those
devices on which apps within each category were used). The number of
devices active within the category is provided in brackets next to each
category name.