Privacy Concerns in Sharing Personal Consumption Data through Online Applications
1. Andreas Kamilaris
Postdoc Researcher IWCMC 2016
Dept. of Computer Science 5-9 September 2016
University of Cyprus
Privacy Concerns in
Sharing Personal Consumption Data
through Online Applications
10. Cyprus Case Study
Total Participants: 198
Male 126 (63.6%)
Female 72 (36.4%)
18-24 62 (31.3%)
25-34 97 (49.0%)
35-49 32 (16.2%)
50-64 7 (3.5%)
1. Online questionnaire (Likert-scale and yes/no questions)
2. Semi-structured mini focus group sessions
Total Participants: 14
Male 8 (57.1%)
Female 6 (42.9%)
18-24 7 (50.0%)
26-32 7 (50.0%)
11. Cyprus Case Study:
Questionnaires
• 85% are willing to share their personal consumption with
their online friends.
• 23% have some concerns about the overall privacy, e.g.
their personal consumption revealed to third parties.
• 60% (of those having concerns) declared willing to share
their consumption only with people they trusted.
• 59% (of those having concerns) were positive to share only
some indicators of their personal consumption.
• 59% would feel more comfortable if the sharing of
consumption data involved street/neighborhood data
(instead of personal consumptions).
12. Reasons for being reluctant:
• Consider those data strictly private.
• Not trusting all of their online contacts.
• Worries of being exposed to thieves.
• Avoid the curiosity of the public about one’s consumption (e.g.
large consumption can be indication of wealth).
• Personal data could be revealed to third parties (e.g. for
advertising).
Specific concerns on the application:
• Display of real-time personal consumption data (high)
• Display of weekly/monthly consumption data (low)
• Showing exact location where they live on the map (medium)
Cyprus Case Study:
Questionnaires
14. Cyprus Case Study: Focus Groups
• “sharing personal data with online friends I trust is OK, as
long as this would happen for a good purpose”
• “common benefits overcome the risks”
• ”the government and many companies know already a lot
about us”
• “banks constitute more serious privacy risks as they know
much about our personal lives”
• “it is our duty to share our consumption, because our
habits affect the environment and the society”
• “sharing only energy savings instead of consumption could
satisfy some privacy concerns”
• “encourage thieves to attempt to steal when the owner is
not there”
• “sometimes people you trust span across multiple
categories”
15. Singapore Case Study
Total Participants: 175
Male 59 (33.7%)
Female 116 (66.3%)
18-24 157 (89.7%)
25-34 15 (10.3%)
• Undergraduate students of the National University of
Singapore.
• The use of Social Electricity assigned as semester project.
• Participation rewarded with a 5% bonus on final grade.
• Results obtained from a final report/assignment prepared by
the students at the end of the semester.
• From the 175 students, 147 filled the final report
• Possibility to participate anonymously.
16. Singapore Case Study: Final Report
• 80% believed the application fully respected their privacy.
• From the rest 20%, 12% believed they exposed some place of
stay information, 4% that users could figure out their real
names and 4% that they exposed their personal consumption.
• 65% were willing to sacrifice their anonymity by using their
real name, considering that this could put some social pressure
on them to reduce their consumption:
• Share with Facebook friends only (5%).
• Share with students from course only (20%).
• Share with students from my tutorial group only (9%).
• Share with particular contacts I can choose from (17%).
• Share with everyone (19 students, 10%).
17. Discussion
• In both studies, a considerable percentage of users (20-23%)
demand for more control over the sharing of their personal
figures, whether this is their consumption or place of stay.
• People reluctant to participate in eco-feedback studies due to
these privacy issues.
• Mechanisms for protecting user’s identity/location and
personal consumption need to be developed.
• Users are asking for application features allowing them of
managing trusted contacts with whom to share sensitive data.
• Sharing indicators of consumption (e.g. daily/weekly savings)
might be more meaningful, with less concerns on privacy.
18. Discussion
• Privacy circles: the Google+ approach of mapping contacts to
categories.
• Some users prefer a one-by-one selection of trustworthy
online contacts and customized circles.
• Online assistants grouping friends according to group based
policy management approaches.
• Automatic mapping of friends to categories based on standard
clustering techniques might be promising.
19. Thank you!
For comments/suggestions/feedback you can email me on
kami@cs.ucy.ac.cy.
Social Electricity App: http://www.social-electricity.com/
SEOP EU Project: http://www.seop-project.eu/
Motivation: online social networking – comparing electricity consumption data
Comparing electricity consumption data raises privacy concerns
A general intro to the SEOP project: mission to develop learning modules, educational content and online eco-feedback platforms, to raise the awareness and knowledge of citizens about energy, the environment and sustainability.
An important achievement of SEOP is Social Electricity, an online application helping people to manage their personal consumption collaboratively, by interacting and comparing with friends, neighbors and other users.
Social Electricity has been awarded the first prize at the 2nd ITU Green ICT Application Challenge.
Two case studies: Cyprus and Singapore
To identify the ”tolerance levels” of our users in regard to sharing electricity data, we asked them whether they would share their consumption figures with people they trusted, and with whom they would be willing to share the consumption of their neighborhood, that of their house as well as the detailed consumption of their household electrical appliances.
The different user categories for sharing were: only me; family members; relatives; close friends; all friends; and everyone.
The graphs are interpreted as follows: starting from ”only me” and ending to ”everyone”, each category is a superset of the previous one.
For example, sharing personal consumption with relatives, this means that the user agrees to share also with his family members and himself.
users have different sharing preferences, dependingonhowpersonalthedataare.While19%arewilling to share their neighborhood’s consumption with everyone, they are reluctant to do this with their home’s consumption, or the consumption of their appliances. In this case, they prefer to share it only with family members, relatives and/or close friends (aggregated 88% in home level and 77% in appliance level). A large percentage (30%) trusts only the other members of their family for sharing their detailed consumption, while only a smaller percentage (14%) wish to share their consumption at appliance level with close friends. Apparently, some users do not trust their close friends, in order to share with them their personal footprint. This percentage is increased in house (36%) and neighborhood level (41%). it is remarkable that from the general to the more specific consumption data, an increasing percentage of users trust only themselves for viewing these values. This percentage starts from 5% in neighborhood level and increases up to 17% in appliance level.
we asked the participants if the proposed user categories for sharing are adequate to them or they wanted to suggest some more. Some users suggested the categories of ”colleagues” (9) and ”business contacts” (6). Many users asked to manually select one-by-one with whom to share their personal data (17) while two users proposed sub-categories below the main categories, for example for selecting only some trustworthy contacts from the list of close friends/relatives.