Widespread web application adoption created a large, complex and varied amount
of data known as Big Data. These data sets have a great value for many economic
and scientific sectors, however they come with additional difficulties when
it comes to storing and analyzing them. Big Data Analytics is the term that
describes the process of researching this massive amount of information in order
to find hidden patterns and correlations. Business Intelligence departments can
now support decision-making processes based on this broad range of data points
collected throughout the lifetime of an application and the designated user’s interaction
with it. However, the abundance and extensive use of Big Data comes
with a number of security and privacy risks that must be addressed. This work
identifies and analyzes these concerns as well as their requirements. Focusing on
user privacy, some of the major issues include: over collection of data in mobile
applications, misuse of data, and multi-source data analysis. These issues can
not always be solved using existing privacy preserving methods. The variety and
velocity of Big Data makes it difficult to distinguish between sensitive and nonsensitive
information, so traditional anonymization techniques can not always be
used. Furthermore, analyzing multi-source datasets can lead to risks of user reidentification.
In this paper we investigate proposed solutions for securing Big
Data as well as ways to maintain data privacy. We look into two major use cases:
healthcare and web analytics, where Big Data is becoming more and more important.
We sum up with a comparison of the requirements and solutions used to
preserve user data privacy for the statistical and clinical data collected in today’s
applications.
