R
elationship between Screen Features of User’s Mobile Terminals
and Cellular Data Usage
Ning Yihe
School of Economics and Management, Beijing University of Posts and Telecommunications, No.10 Xitucheng Road,
Haidian District, Beijing, China
Keywords: screen size, mobile terminal, cellular data consumption, telecommunications company.
Abstract: The cell-phone has been standard equipment for almost everyone in modern society, and mobile data
services have gradually become the major business for global telecommunications companies. This study
aims to conduct research on the relationship between screen features of user’s mobile terminals and their
cellular data usage, we established an individual level regression model to explore the correlation
relationship between screen features of user’s mobile terminals and cellular data usage based on 50,000
records of consumers’ cellular data consumption and their basic information provided by
Telecommunications Corporation in China. During the study, we carried out regression analysis from four
perspectives, which are all terminals, non-smart phones, smart phones and tablets. It is firstly observed that
there is an overall positive and significant correlation between screen features of user’s mobile terminals
and cellular data usage over the entire range of screen sizes. Secondly, we also found that the effects of
user’s age on cellular data consumptions are always significantly negative for different terminal datasets,
that is to say, the older the user is, the less cellular data the user use. This study has not only provided some
suggestions for telecommunications companies, but also made reference recommendations for mobile
terminal manufacturers.
1 INTRODUCTION
With the development of communication
technology, people have entered the era of mobile
Internet, in which the cell-phone has been standard
equipment for almost everyone in modern society.
Making a general observation of the global market,
mobile data service has become one of the main
services provided by global telecom operators for
users (Olla et al.,2002, Lu et al.,2008). Cisco Visual
Networking Index: Global Mobile Data Traffic
Forecast Update: 2014–2019, issued by Cisco in
February 3, 2015, indicated the facts that the global
mobile data traffic had increased by 64% in 2014, it
had increased from 1.5 exabytes per month in 2013
to 2.5 exabytes per month in 2014(Cisco,2015).
Regarding different consumers using different
mobile terminals with various screen sizes, there are
also significant discrepancies on consumers’ cellular
data consumption. So, what is the relationship
between screen features of users’ mobile terminals
and users’ cellular data usage? This study aims to
conduct research on the relationship between screen
features of users’ mobile terminals and their cellular
data usage.
The changes of mobile terminal drive the
development of Internet, the Mobile Internet is
evolving, and the study of mobile terminals and
Mobile Internet has become a new research
direction. Adipat, Zhang, Zhou et al. studied the
effects of browsing the web in the form of TreeView
on handheld devices (Adipat et al.,2011). According
to the influences of demonstration adaptation
method, they proposed a mixed method to adjust the
rending effects of web page, this method improves
notably the Mobile Web experiences users perceived
and the experiences of web browsing, an important
factor considered in this study is the screen size of
mobile terminal. Ghose et al. explored the changes
of user’s browsing behavior in mobile terminal and
PC in the Internet era, their findings show that when
the screen sizes of mobile terminal is small, the costs
of user searching information will increase, and
31
Yihe N.
Relationship between Screen Features of Userâ
˘
A
´
Zs Mobile Terminals and Cellular Data Usage.
DOI: 10.5220/0006443200310035
In ISME 2016 - Information Science and Management Engineering IV (ISME 2016), pages 31-35
ISBN: 978-989-758-208-0
Copyright
c
 2016 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
31

users are more sensitive to the position of
information in mobile terminal, users are more likely
to click on the link at the top of the cellphone screen,
so the effects of ranking are better in mobile
terminal (Ghose et al.,2011, Ghose et al.,2013,
Ghose et al.,2014). However, at this stage there is no
direct the academic study indicates that different
screen sizes will affect the usage of mobile data
traffic, and we can conclude from the previous
researches that the changes of mobile terminal can
bring large impact on user’s behaviors. Making
research on the relationship between the screen
features of user’s mobile terminals and cellular data
usage can fill the research gap.
Based on this background, this study aims to
conduct research on the relationship between screen
features of user’s mobile terminals and their cellular
data usage by utilizing the massive amounts of
consumers’ data provided by a telecommunications
company. Specifically, we established an individual
level regression model to explore the correlation
relationship between screen features of user’s
mobile terminals and cellular data usage based on
50,000 records of consumers’ cellular data
consumption and their basic information provided
by Telecommunications Corporation in China.
This study has great value on the terminal
manufacturers and telecom operators, exploring the
hidden relations has both theoretical and practical
significance for users, terminal manufacturers and
telecom operators. In theory, drawing on and using a
large number of discipline research methods and
findings in the study, exploring the relationship
between the screen features of user’s mobile
terminals and their mobile data usage based on users
data, have an important significance on data mining
and data analysis in the context of big data. From a
practical point of view, on one hand, this study
provide practical cases of data analysis and
application for the telecommunications industry who
have huge amounts of user data, it can help telecom
operators understand users’ online behaviors better,
and pricing products combined the properties of
terminals offers a new perspective for telecom
operators. On the other hand, for mobile terminal
equipment manufacturers, this study can help mobile
terminal equipment manufacturers choose more
suitable screen sizes of mobile terminals, make the
user experience better.
2 METHOD
2.1 Multiple Linear Regression
In this study, we employ a multiple linear regression
model to explore the relationship between the screen
size of mobile terminal and the mobile data traffic
from the user level. We can find from the descriptive
statistics that the screen size of mobile terminal does
have influence on user’s Cellular data usage, and
user’s personal characteristics, such as user’s age,
gender and rank, also have a certain influence on
user’s Cellular data usage, so we also include the
personal characteristics of user into the variables that
affect user’s cellular data usage. Taking into account
the brand of mobile terminal will also affect the way
customer use the cellphone, and further affect user’s
Cellular data usage, so we choose the personal
characteristics of user, the screen size of mobile
terminal and the brand of mobile terminal as the
independent variables, the Cellular data usage user
used as the dependent variable. The screen size is
the primary factor, and user’s personal
characteristics and the terminal brand as the control
variables, the employed regressive model as be
expressed as follows:
(1)
where y is the dependent variable, represents the
cellular data traffic, and its unit is MB; i represent
the user, Screen
i
is the screen size of the mobile
terminal that ith user own; Sex
i
is the gender of ith
user, 1 represents female and 0 represents male; Age
i
the age of ith user; Diamond
i
, Gold
i
, Silver
i
are three
dummy variables representing user rank, each user
only belongs to one rank of these three ranks; Brand
is the dummy variable of brand, it can be used to
evaluate the influences of the mainstream cell-phone
brands on cellular data usage. The error term ε is
used to adjust error brought by the prediction model.
2.2 Data
The data used in our study is provided by a telecom
company in China, which record 50,000 consumers’
information including cellular data consumption and
their basic information in April 2014. The raw data
provided by the operator only records the terminal
model, through crawler technology obtaining the
attribute information of terminals, and then relating
it to the screen size customers used, we acquired our
final data. In our dataset, the total data recordings
amount to 48890 items, including the tablets
ii i i
ii
y
Screen Sex Age Diamond
Gold Silver Brand ε
=
++ +
++ + +
ISME 2016 - Information Science and Management Engineering IV
32
ISME 2016 - International Conference on Information System and Management Engineering
32

recordings 1191 items, the fields is listed as follows
： 1)user_id, unique identification for identifying
each user; 2)sex, representing the gender of user,M
is male and F is female; 3) age, which is calculated
based on user’s birthday; 4) gprs_stream,
representing the consumption of cellular data, it
records the total amount of cellular data user used in
April 2014, and its unit is MB; 5) screen, recording
the screen size of mobile terminal; 6) VIP Class,
there are three user ranks, Silver, Gold, Diamond,
each user belongs to one and only one rank; 7)
brand, which is a dummy variable and represents the
effects of terminal brand, main brands of terminal
equipment contain Apple, Samsung, Nokia, Xiaomi,
Huawei and so on.
According to the terminal type, we divide the
users into phone group and tablet group, and we
conducted a descriptive statistical analysis. In the
phone group, 19.86% of total users are female, and
male accounts for 80.14%. The average screen size
of user’s mobile phone is 4.2 inches, and the average
age of user is 31 year old, in addition, the average
cellular data consumption of user is 264.52 MB in
current month. In the tablets group, female accounts
for 20.74% while the proportion of male is 79.26%.
The average screen size of user’s tablet is 8.7 inches,
and the average age of user is 36 years old, which is
slightly older than the average age of phone group.
Besides, the average cellular data consumption of
user is 595.75 MB in current month. By crossing
analysis, we found that the screen size of user’s
terminal, personal characteristics and brands have an
impact on user’s cellular data consumption.
3 EMPIRICAL RESULTS
In this paper, we carry on linear regressive analysis
based on the overall sample data, mobile terminal
sample data and tablet computer sample data
respectively, and the empirical results are shown in
Table 1. This study carried on regressive analysis
from the user personal level and equipment level, as
is shown in Table 1, the first column is the
dependent variables in the model, and the first row is
the terminal equipment. From the equipment level,
we divide terminal into overall terminals (that is,
cell-phone & tablet computer), cell-phone under 3
inches, cell-phone above 3 inches and tablet
computer. We make 3 inches as a cut-off point for
the screen size of cell-phone, since the screen size of
smart-phone is all above 3 inches, and that of most
non smart-phones is less than 3 inches. Non smart-
phone can only access to the Internet in textual way
and consumes little cellular data, so the smart-phone
is the focus of our research.
The second column in Table 1 presents the
regression results for all equipment. It can be found
that for the overall terminals including cell-phone
and tablet computer, screen size, sex, age, user rank
and brand are all significant to user’s cellular data
usage at the 5% level. The effect of screen size on
cellular data usage is positive, and its standard error
is quite small. With respect to user’s personal
characteristics, the effects of two factors, sex and
age, is negative. That is to say, women’s cellular
data usages are less than men, and the bigger user’s
age is, the more cellular data he consumed. User
rank also affects the cellular data usage significantly,
and the effect is positive, but the standard error is
large (65.647 for the SE of Diamond rank). There is
a significantly positive correlation between terminal
brand and cellular data usage.
The third column in Table 1 presents the
regression results for cell-phone under 3 inches
which is non smart-phone. Because there is no
Diamond rank user among users who use cell-phone
under 3 inches, so the corresponding term is vacant.
As is shown in the table, for non-smart phone,
screen size, age and user rank of Silver are
significant to user’s cellular data usage at the 5%
level, but the effects of user’s sex, brand and user
rank of Gold are not significant. The effect of screen
size on cellular data usage is negative significantly,
which means that for non smart-phone users, the
bigger the screen size of mobile phone is, the less
the cellular data they consume. There is also a
significantly negative correlation between user’s age
and cellular data usage, as users grow older, their
cellular data usage have reduced. The effect of
Silver rank on cellular data usage is significantly
positive, but the error is large.The fourth column in
Table 1 presents the regression results for cell-phone
more 3 inches which is smart-phone. Because there
is no Gold and Silver rank user among users who
use cell-phone above 3 inches, so the corresponding
terms are vacant. As is shown in the table, for smart-
phone, screen size, age, brand and user rank of
Diamond are significant to user’s cellular data usage
at the 5% level, but the effect of sex is not
significant. The effect of screen size on cellular data
usage is significantly positive, and its error is quite
small, which means for smart-phone users, the
bigger the screen size of mobile phone is, the more
the cellular data they consume. There is a
significantly negative correlation between user’s age
and cellular data usage, as users grow older, their
cellular data usage have reduced. The effects of
Relationship between Screen Features of User’s Mobile Terminals and Cellular Data Usage
33
Relationship between Screen Features of Userâ
˘
A
´
Zs Mobile Terminals and Cellular Data Usage
33

brand and user rank on cellular data usage is
significantly positive, but the error of Silver rank is
large.
Table 1: Results of the multiple linear regression model
Notes: *: p-value<0.1, **: p-value<0.05, ***: p-value<0.01.
The fifth column in Table 1 presents the
regression results for tablet computer. As is shown in
the table, for tablet computer, screen size, age are
significant to user’s cellular data usage at the 5%
level, but the effects of sex, user rank, and terminal
brand are not significant. The effect of screen size on
cellular data usage is significantly positive, which
means for tablet computer users, the bigger the
screen size of mobile phone is, the more the cellular
data they consume. There is a significantly negative
correlation between user’s age and cellular data
usage, as users grow older, their cellular data usage
have reduced.
As a whole, no matter how to categorize
equipment, the effects of screen size on user’s
cellular data usage are always significant, especially
for smart-phone and tablet computer, of which the
screen size effect is positive. And the effects of
user’s age are always significantly negative for
different terminal datasets, that is to say, the older
the user is, the less cellular data the user use. As for
other variables, the effects and significance are
various for different datasets.
4 CONCLUSIONS
In the paper, we carried out regression analysis to
explore the relationship between screen features of
user’s mobile terminals and their cellular data usage
by utilizing the massive amounts of consumers’ data
provided by Telecommunications Corporation in
China. The study is conducted from four
perspectives, which are all terminals, non-smart
phones, smart phones and tablets. Several
conclusions can be drawn from the empirical results,
firstly, there is an overall positive and significant
correlation between screen features of user’s mobile
terminals and cellular data usage over the entire
range of screen sizes. Secondly, we also found that
the effects of user’s age on cellular data
consumptions are always significantly negative for
different terminal datasets, that is to say, the older
the user is, the less cellular data the user use.
According to the results of this study, we
propose some recommendations for products and
pricing strategies of telecommunications companies.
First, telecommunication enterprises can establish a
more flexible pricing standards for cellular data, and
provide a stepped pricing method, which can
stimulate customers to consume more cellular data.
Second, telecommunication enterprises can
strengthen competition advantage via adopting
differentiation strategies, which can set different
prices for different groups of consumers. For
consumers aged from 20 to 40 years old, telecom
firms should design a postage service package,
which contains a wealth of cellular data and a low
price, to better meet the needs of this part of the
young and middle-aged population. Third, telecom
firms can optimize package services configuration
and charge combined the screen size of consumer’s
terminal. For terminals with different screen sizes,
they can offer different contract-phone with
corresponding data program, which can better cater
cell-phone & tablet
computer
cell-phone under 3
inches
cell-phone above 3
inches
tablet computer
Screen
i
43.710***
（2.868）
-72.500**
（31.038）
5.724***
（4.463）
84.926***
（28.253）
Sex
i
-29.285***
（7.287）
35.583
（33.652）
-30.318
（7.484）
-26.293
（76.261）
Age
i
-8.319***
（0.360）
-5.509***
（1.302）
-6.749***
（0.373）
-7.449**
（3.521）
Diamond
i
753.577***
（65.647）
—
705.466***
（66.122）
-493.115
（1057.303）
Gold
i
608.986***
（25.031）
61.563
（321.781）
—
-166.980
（611.609）
Silver
i
325.944***
（9.479）
120.547**
（55.400）
—
231.253
（180.202）
Brand
88.294***
（7.853）
-21.662
（32.855）
128.233***
（7.688）
324.476
（435.124）
R
2
0.049 0.023 0.015 0.012
ISME 2016 - Information Science and Management Engineering IV
34
ISME 2016 - International Conference on Information System and Management Engineering
34

to the market and meet the needs of consumers,
obtaining steady high-value users eventually.
Our study found consumers preferred screen
sizes of mobile device through an in-depth research
based on consumers’ data, which is valuable for
mobile terminal manufacturers. According to the
results of our study, we present useful proposals for
terminal manufacturers. First, mobile terminal
manufacturers should meet the needs of majority
group in the design of terminal’s screen sizes, and
screen size should not be too big or too small.
Second, the tablet device manufacturers can use
current user's choice on screen size as reference,
setting tablet screen size at 7.9 inches or 9.7 inches
which consumers prefer. Third, the terminal
manufacturers should cooperate with
telecommunication companies to design contract
mobile phones to attract consumers that contains
medium screen size and high cellular data, reaching
a win-win result in the future.
REFERENCES
Adipat B, Zhang D, Zhou L. J., 2011. THE EFFECTS OF
TREE-VIEW BASED PRESENTATION ADAPTATION
ON MOBILE WEB BROWSING. Mis Quarterly,
35(1):99-121
Cisco, J., 2015. Visual Networking Index: Global Mobile
Data Traffic Forecast Update:2014–2019.
Ghose A, Han S P., J., 2011. An Empirical Analysis of
User Content Generation and Usage Behavior on the
Mobile Internet. Management Science, 57(9):1671-
1691
Ghose A, Goldfarb A, Han S P, J., 2013. How Is the
Mobile Internet Different? Search Costs and Local
Activities. Information Systems Research, 24(3):613-
631.
Ghose A, Han S P., J., 2014. Estimating demand for
mobile applications in the new economy. Management
Science Journal of the Institute for Operations
Research & the Management Sciences, 60(6): 1470-
1488.
Lu J, Liu C, Yu C S, et al, J., 2008. Determinants of
accepting wireless mobile data services in China.
Information & Management, 45(1): 52-64.
Olla P, Patel N V, J., 2002. A value chain model for mobile
data service providers. Telecommunications Policy,
26(9): 551-571.
Relationship between Screen Features of User’s Mobile Terminals and Cellular Data Usage
35
Relationship between Screen Features of Userâ
˘
A
´
Zs Mobile Terminals and Cellular Data Usage
35