RELATIONS BETWEEN LEARNING ANALYTICS AND DATA PRIVACY IN MOOCs

RELATIONS BETWEEN LEARNING ANALYTICS AND
DATA PRIVACY IN MOOCs
Malinka Ivanova
Technical University of Sofia, Bulgaria
Carmen Holotescu
„Ioan Slavici” University of Timisoara, Romania
Gabriela Grosseck
West University of Timisoara, Romania
Cătălin Iapă
University Politehnica Timisoara, Romania
eLearning and Software for Education
Conference - eLSE 2016, Workshop on Open
Educational Resources and MOOC,
Bucharest, 21-22 April 2016

AIM
To summarize and analyse the current projects and
practices, best learning cases and research
standpoints in the area of MOOCs data use for
learning analytics purposes and at this base a model
presenting a possibility to ensure data privacy in
MOOCs to be developed

OUTLINE
Introduction and problem definition
Research method
Collected data by MOOCs platform
Big data and MOOCs analytics
CIC data privacy model in MOOCs
Conclusions

understanding and
optimizing learning
and the
environments in
which it occurs
Learning
Analytics
-measurement,
-collection,
-analysis and
-reporting of
data about
learners and
their contexts
is defined as for purposes of
Siemens, Long, 2011
-LMS
-Social media platforms
-MOOCs
-Online
-Blended-learning
courses

Data about learners
Learning support
Records of students’ activity in LMSs
•logging, posting and commenting messages, accessing
materials, posting assignments
Results in previous courses
Preferences
Learning optimization and
understanding
Prediction and improvement
students’ academic performance
Helping students „at risk” with
prompt feedback
Learning
Analytics

METHOD
Review of
privacy policy of five
MOOCs platforms
Review of
started projects and
published scientific
papers
Research questions
What kind of students’ private data are
shared and why?
What are the privacy models provided
by MOOCs?
What data are needed for Learning
Analytics modules?
Are there any third party actors
processing the students’ data and why?
Which are the relations between
Learning Analytics and Data Privacy?
Is it possible the learners to be
protected?
Development
of a model reflecting
the data privacy in
MOOCs

COLLECTED DATA BY MOOCs PLATFORMS AND THIRD PARTIES
The term:
• Personal data is legally recognized and it is defined as “any information relating
to an identified or identifiable natural person” (Directive 95/46/EC)
• Private data is not used by juridical science, but it is often utilized in other fields
of science – for example in computer science
• Sensitive data - special category of data in legal documents related to “racial or
ethnic origin, political opinions, religious or philosophical beliefs, trade-union
membership, and the processing of data concerning health or sex life” (Directive
95/46/EC)
• Confidential data - used only between two parties, keeping it in secret; it is not
available for public use
Personal data Non-personal data
Private data
Sensitive data
Confidential data
Continuum of collected data for research and educational purposes

Factors/
MOOCs
Platforms
edX Coursera FutureLearn CanvasNetwork Open2Study
Collected
information
by
platforms
Personal
information about:
-sign up;
-participation in online
courses;
-registration for a paid
certificate;
-at sending email
messages;
-participation in public
forums;
-student learning
performance;
-which, when, where
pages are visited;
-which hyperlinks and
other user interface
controls are used.
A.Personal
information:
-at registration; -for
user account update;
-to purchase products or
services;
-to complete surveys;
-to sign-up for email
updates;
-to participate in
forums;
-to send email;
-to participate in online
courses.
B.Non-personal
information:
- which, when, where
pages are visited;
-which hyperlinks are
visited;
-URLs from which
Cursera is visited;
-log of IP address, OS
Personal information:
-for access to website;
-to use online courses and
content;
-to register or post notes,
assignments or other
material;
-information when user
report a problem;
-location, gender and
educational history or
qualifications;
-other information for
delivering the best service;
-results of assessments;
–at purchase a paid service -
credit card or other payment
details;
- a record of email
correspondence;
-data about usage of
content;
-IP address.
Personal information:
-data for login;
-messages and stored files;
-email content;
-survey data;
- how user interacts with
applications;
- encountered errors by users
when use platform;
- device identifiers;
-how often users visit the
site, what pages are visited,
what other sites are used for
coming to the site;
-IP address, operating system,
browser type, domain name.
A.Personal information:
-at registration;
-educational qualifications;
-academic results;
-banking and payment details;
-tax file number;
- responses to surveys;
-how, when and why services
are used.
B. Sensitive information
-language background;
-citizenship;
-disability;
-health information.
C.Non-personal
information
- website visit - date, time and
duration, search terms,
viewed pages;
-IP address, type of device,
browser and OS.

Collected
financial
information
By third party
payment site
-A. By third party
payment site;
-B. By Coursera:
credit card
information
By FutureLearn: data
about credit card or
other payment
details at purchasing
a paid service
- By Open2Study that
uses EFTPOS and
online technologies
for payment
Factors/
MOOCs
Platforms

Usage of
collected
information
by
platforms
-to personalize
the course
material;
-to improve
learning
process.
A.Personal
information for:
-delivering specific
courses and/or
services;
-sending updates
about online
courses and
events.
B.Non-Personal
information for:
-improvement of
services;
-other business
purposes.
- to improve,
maintain and protect
the quality of web
site, online courses
and content;
-to personalize
browsing;
-for announcements
and email
notifications;
-for opinion gathering
- to create and
maintain user
account;
-to provide better
services;
- to personalize and
improve learner
experience;
- to send
administrative e-mail
and announcement;
-to send surveys
- for user
identification and
verification;
-for communication;
- for delivering of
educational services;
- for improvement of
web site services
Factors/
MOOCs
Platforms

Usage of
collected
information
from third
parties
-by educational
institutions;
-by other
providers of
courses in edX
- by service
providers,
vendors and
contractors;
-by university
partners and
other business
partners;
- by government
authorities
- by FutureLearn
partners;
-by course and
content providers;
-other third parties
- third party service
providers;
- for merger, financing,
acquisition,
bankruptcy,
dissolution,
transaction, sale
purposes
-by Australian
government
- by organizations
administrating the
business;
- by financial
institutions;
-by service
providers and
research agencies,
mailing houses,
postal, freight and
courier service
providers, printers
and distributors of
direct marketing
material, external
business advisers
Factors/
MOOCs
Platforms

Security -software
programme for
data protection
through
administrative,
physical, and
technical
safeguards
-industry
standard: physical,
technical and
administrative
security measures
- not sharing
information with
third parties,
except before
mentioned
-technical and
organizational
security measures
-secure web site
- protect against
unauthorized access,
information use, or
disclosure
-any posted content
using CanvasNetwork
services is at user risk
-ICT security;
-secure office
access;
-personnel security
and training;
-workplace policies
Factors/
MOOCs
Platforms

Usage of
cookies
cookies: for
collecting IP
address,
operating
system,
and browser
information
cookies and/or
web beacons:
-to identify users;
-to personalize
experience;
-to identify repeat
visitors;
-to determine the
type of content
and spending time
cookies:
- to enhance learner
experience;
-for better
understand how
learners use the
website;
- cookies from third
party social media
websites
-cookies and web
beacons:
for information about
user visit and
searched/ viewed
content
- flash cookies: to
store user preferences
and personalize visits
cookies:
-to hold anonymous
session;
- to personalize
website visits
Factors/
MOOCs
Platforms

Privacy
policy
-Own edX
privacy policy;
-Family
Educational
Rights and
Privacy (FERPA)
for education
records
Own Cursera
privacy policy
Own FutureLearn
privacy policy
Own CanvasNetwork
privacy policy
-Own complying
with:
the Privacy Act
1988, State and
Territory health
privacy legislation,
the Spam Act 2003,
the Do Not Call
Register Act 2006
Factors/
MOOCs
Platforms

BIG DATA AND MOOCs ANALYTICS
• Report about new technology approaches and challenges for
performance of MOOCs analytics (O’Reilly, Veeramachaneni, 2014)
• Emerging of web platforms facilitating huge groups to contribute in
development of analytics
MOOCviz – platform for collaborative
MOOC visualizations
FeatureFactory – interactive platform
for presenting prediction problems
LabelMe-Text – platform designed in
support of annotating in forum posts

• Report of LASyM learning analytics system for MOOCs:
 mines “big data”
 analyzes learning outcomes and assessment of learners
 delivers information that could be used for design of optimized MOOCs
• The aim of LASyM is to minimize “at-risk” MOOC students through
their identification at the earliest possible stage
Tabaa, Medouri, 2013

• Ascertainment about the growing interest among researchers and
educators to open analytical systems like edX Insights and Tin Can
• Their usage could improve effectiveness of course design in different
subjects:
 different learning scenarios could be developed for studying computer
science and art history
 improvement of effectiveness of peer ranking and improvement of course
assignments
Godwin-Jones, 2014

• European project Multiple MOOC Aggregator (EMMA) reports
development of a MOOC platform that aggregates existing European
MOOC courses with possibility for learners to design their own
personal learning environment
• EMMA platform possesses analytical application for ensuring
realization of personal learning paths according to individual learning
needs
• The learning process is monitored, achievements are controlled and
development of a competence like “learning to learn” is improved
Brouns, Tammets, Padrón-Nápoles, 2014

• European project EDSA (European Data Science Academy) - a data
mining process in MOOCs on Coursera is applied with aim the
learning process to be analyzed
• An experiment with students from Eindhoven University of
Technology registered for participation in MOOCs on Coursera is
performed
 their behavior is analyzed
 a learning model of successful and failed students is created
EDSA project, http://edsa-project.eu/resources/learning-analytics

CIC DATA PRIVACY MODEL IN MOOCs
• Importance of applying learning
analytics to “big data” mining for
improvement of MOOCs learning
scenarios
• A huge massive of information
concerning any learner is utilized
and personal data are not kept
private
• The model CIC (Classification-
Information-Choice) describes six
measures for ensuring data
privacy in scenarios of large scale
learning
Classification. Appropriate
classification of collected personal data/
non-personal data in groups and
classification of third parties
Information. Preparation of
informative tools informing learners
about type of collected data,
mechanisms for collecting and storing,
data usage and possibilities for choices
Choice. Choices regarding data sharing
to MOOCs platform, data giving to third
parties, for movement personal and non-
personal data from one classification
group to other

CIC DATA
PRIVACY
MODEL IN
MOOCs

CONCLUSIONS
• Intensive work on data privacy in different applicable areas in
practical and theoretical aspect, including in MOOCs learning
• Anyway, this topic is still in its immature form and it is a need for
further development of policies, mechanisms and tools
• In the existing documents the privacy of personal data is well
discussed, but nothing is mentioned about the privacy of non-
personal data
• As it is seen a connection between personal and non-personal data
exists and this connection should be taken into consideration when a
data privacy is forming

CONCLUSIONS
• The created CIC model reflects on the current situation:
 there are privacy policies written for a given MOOCs platform, but they are
not complete;
 existing learning analytics tools are utilized with minimum care for data
privacy;
 “big data” is collected and stored in unsafe digital repositories
• It reveals the big picture for undertaking measures for improvement
of data privacy in MOOCs
• It could be used as a recommendation tool guiding developers of
MOOCs in realization of more complete data privacy

Thank you!
• Picture is taken from:
http://cdn.ttgtmedia.com/visuals/ComputerWeekly/Hero%20Images/data-
privacy-fotolia.jpg

RELATIONS BETWEEN LEARNING ANALYTICS AND DATA PRIVACY IN MOOCs

More Related Content

Viewers also liked (7)

Similar to RELATIONS BETWEEN LEARNING ANALYTICS AND DATA PRIVACY IN MOOCs (20)

More from Malinka Ivanova (20)

Recently uploaded (20)

RELATIONS BETWEEN LEARNING ANALYTICS AND DATA PRIVACY IN MOOCs