Security and Privacy Issues in Analytics

This report discusses the security and privacy issues faced in analytics, including data security measures and protection practices. It also explores the ethical concerns in building an image recognition system for law enforcement.

Running head: SECURITY AND PRIVACY ISSUES IN ANALYTICS
Security and Privacy Issues in Analytics
Name of the Student
Name of the University
Author’s Note:

1
SECURITY AND PRIVACY ISSUES IN ANALYTICS
Table of Contents
1. Executive Summary...............................................................................................................2
2. Introduction............................................................................................................................3
3. Discussion..............................................................................................................................3
3.1 Privacy Issues Raised by Analytical Datasets..................................................................3
3.2 Ethical Issues and Analytics.............................................................................................6
4. Conclusion..............................................................................................................................9
References................................................................................................................................10

2
SECURITY AND PRIVACY ISSUES IN ANALYTICS
1. Executive Summary
The main aim of this report is to know about the security and privacy issues faced during
dealing with the data and analytics. Data security is termed as the core and significant
procedure for protection of data from any type of authorized and unauthenticated access as
well as corruption of data in the entire life cycle. This type of data security eventually
involves tokenization, data encryption and major management practices, which are
responsible for protecting the data within all types of platforms and applications. The
organizations in the entire world are eventually investing heavily within the IT or information
technology cyber defence capabilities for the core purpose of protecting and securing the
critical assets. When an enterprise requires to protect their brand, customer information and
intellectual capitals, it subsequently refers to the fact that incident detection and responses are
important to protect the organizational interests with three common elements of technology,
processes and people. Security and privacy issues are also common for data analytics and the
sensitive information often becomes quite vulnerable to security. This report has properly
described about the several privacy issues that are being raised by the analytical datasets in
the Netflix data challenge, involvement of technical issues as well as ethical issues, which
could arise from utilization of machine learning.

3
SECURITY AND PRIVACY ISSUES IN ANALYTICS
2. Introduction
The analytical datasets should be protected for ensuring that the network traffic and
business applications are secured (Dwork and Roth 2014). The following report outlines a
brief discussion on the security, privacy and ethical issues in analytics with relevant details.
3. Discussion
3.1 Privacy Issues Raised by Analytical Datasets
The sparse data refers to the data there are several gaps present within the data that is
being recorded. These types of data come from sensors or any other non-information
technology based devices. This sparse data almost goes in one way from sensor to the
network. There are several devices that are required to be analysed under every condition
(Cao et al. 2014). The new data scientist of the tech start-up has got the distinctive
responsibility to understand the Netflix data challenge and various privacy issues that could
be possible for the organization. Moreover, it was observed that he entire data challenge
occurred due to sparse data and analytical dataset security.
There are some of the most significant technologies that are required to be considered
while dealing with these privacy issues like disk encryption, software based mechanisms,
backups, data erasure and many more (Zhu et al. 2015). The non information contextual data
are being considered for anomaly detection logic. The technologies analyse and accumulate
real time data, which involves asset metadata, threat intelligence and IP context. All of such
forms of data could be utilized for both immediate investigations and threat response. The
CTO of the organization is considering about the data threat or issue faced by Netflix. A high
dimensional sparse dataset is responsible for analysing the sparse data and storing that data
securely.

4
SECURITY AND PRIVACY ISSUES IN ANALYTICS
The most significant privacy vulnerabilities that are generally faced while securing
sparse dataset are as follows:
i) Protection of Transaction Data and Logs: The first and the most important privacy
vulnerability that is being faced while securing the sparse dataset is protection of transaction
data and logs. Since, Netflix has to deal with several users’ data, it is extremely vital for them
to protect their transactional data and logs. For this purpose one has to break the anonymity
of the Netflix Prize Dataset (Uhlerop, Slavković and Fienberg 2013). The anonymity can be
broken after considering few significant steps and it is required to protect the dataset under
every circumstance. When the data size is incremented, the availability and scalability makes
auto tier important for the management of data storage. The storage location is required to be
secured and hence Netflix data can be reduced.
ii) Validation and Filtration of End Point Inputs: Another important and noteworthy
privacy and security issue that is required to be considered for sparse dataset anonymity, is
the significant validation and filtration of the end point inputs (Chen et al. 2013). The end
point devices are the major factors to maintain sparse dataset. The storage, processing as well
as required works are being performed with an input data that is eventually provided for
every end point. Hence, Netflix should ensure to utilize a legitimate and authenticated end
point device. The validation and filtration of the end point inputs is required for
understanding the large volumes of data and datasets (Lyu, Su and Li 2017). The reason for
these types of breaches are the majority of security application, which is designed for storing
certain amount of sparse data and the respective security technologies could become
inefficient in managing the dynamic data.
iii) Securing Distributed Framework Calculations and Processes: The third type of
distinct and subsequent privacy or security issue that is extremely vulnerable for the sparse

5
SECURITY AND PRIVACY ISSUES IN ANALYTICS
dataset and Netflix anonymity is lack of securing a distributed framework calculations and
any other process (Blum and Roth 2013). The computational security as well as the other
digitalized assets within the distributed framework in Netflix prized dataset eventually lack
any type of security protection. The two major preventions for this type of data security are
protecting the dataset and securing the mappers within the presence of an unauthenticated and
unauthorized mapper.
iv) Protection and Security of the Data in Real Time: The next significant and
distinctive privacy and security issue that is required to be eradication under every
circumstance is the lack of protection and security of real time data. For the large amount of
this data generation, Netflix is often unable to preserve the regular data check. However, it is
extremely beneficial to accomplish such type of security check as well as observations within
the real time or even almost within real time.
v) Lack of Granular Access Control: This is again one of the major security issue
that is to be considered for Netflix prized dataset. The granular access of the sparse data are
stored by NoSQL databases with a mandatory access control and stronger authentication
process. The lack of granular auditing is yet another significant issue that is required to be
eradicated on time for the sparse data. Analysis of several types of logs can be beneficial and
this data can be helpful for recognizing the type of malicious activities and cyber-attacks.
The anonymity of the sparse training data of Netflix could attacked if the above
mentioned issues are not removed on time (Greenland, Mansournia and Altman 2016).
Moreover, Netflix falls under the grouping of high dimensional sparse dataset and additional
data of number of subscribers of Netflix is required. De-anonymization of the large sparse
datasets is possible with database model, sparsity and similarity. After experiment, it is
observed that this Netflix prize dataset is highly sparse and for these vast collection of

6
SECURITY AND PRIVACY ISSUES IN ANALYTICS
records, there is not one single record with the subsequent similarity score more than 0.5
within the complete 500000 recorded dataset when the sets of movies are considered without
considering the numerical ratings. Releasing and sanitation of data is required for
implementing the changes.
3.2 Ethical Issues and Analytics
The CTO of the tech start-up is thinking about bidding for a contract with the local
government for the purpose of building an image recognition system for the law enforcement
(Mason 2017). A face recognition system is the most distinct technology that has the
capability to identify or verify an individual from a video frame or digital image. There are
several distinct methods by which the facial recognition systems can work and the work is
completed by comparing the selected facial features from a provided image with faces in the
database. This is even termed as a biometric AI based application, which could easily and
promptly recognize a person after analysing the patterns on the basis of an individual’s facial
shape and textures (Slade and Prinsloo 2013). Since, CTO of his start-up has thought of
bidding for a contract with the local government to build up this system, it is extremely
important and significant to consider the ethical issues for such a contract.
The major and the most significant ethical issues that are required to be considered
here are as follows:
i) Data Theft: This is the first and the most significant ethical issue that should be
considered for bidding of this contract is theft of data. There are several chances of data
hacking and the data could even be forged by the hackers. If the data is present within the
cloud, indexed by Google, it is likely to get hacked by the hackers and data might lose
confidentiality eventually (Mealer and Jones 2014). However, one cannot change someone’s
face and there is an advantage as well that the cloud never forgets any face. Facial

7
SECURITY AND PRIVACY ISSUES IN ANALYTICS
recognition system is also effective for solving crime for perpetuating social stigma, ethnic
and racial profiling and many more.
ii) Terrorist Attacks: Since, the start-up will be starting a contract with the local
government, it is required to consider the ethical issues faced by the government people.
There is always a high chance the terrorist attacks should be considered and the facial
recognition system would have a chance to identify the issues and complexities to a higher
level (Miller and Blackler 2017). These types of attacks are extremely significant to identify
the terrorists of a nation and a database of suspected criminals would have the ability to
recognize the terrorists.
iii) Smart Closed Circuit Television: From the ethics based point of view, the smart
closed circuit television is extremely interesting as it includes two distinct contested
technologies of biometrics and video technology. FaceIt is the software engine, which is
being run on a system for detecting and recognizing the human faces (Slade and Prinsloo
2013). It eventually undertakes human faces as the inputs after encoding them into digital
images. This particular software is extremely ethical and has the capability of identifying the
human faces without much complexities.
iv) Maintenance of Reasonable Data Security Protections: Another important and
significant ethical issue that is required to be considered for the facial recognition system is
that the chief technical officer of this start-up should consider the maintenance of reasonable
data security protections (Martin 2015). For the consumer’s images, this type of protection is
highly mandatory and they should store these images after putting protections in place and
hence preventing any kind of unauthorized or unauthenticated scraping that leads to the
unintended secondary utilizations.

8
SECURITY AND PRIVACY ISSUES IN ANALYTICS
v) Uses of Digital Sign: Another significant and noteworthy ethical issue that is
required to be considered for the facial recognition system for this contract between
government and the tech start-up is utilization of digital signs (Grinbaum and Groves 2013).
This type of digitalized signs comprise of the significant capability to detect demography.
The digital signs should give clearer notice to the clients that such technologies are in use,
even before the clients are coming into contact with the respective signs. These digital signs
can easily be used in the health care sector as it helps in receiving new sets of images of the
various patients, who have any kind of genetic disorders. For maintenance of ethics and trust
with the patients, the tech start-up must consider the involvement of relevant community
stakeholders for the implementation of this facial recognition system and decisions for
establishment and improvement of practices for informing patients regarding the utilization
of this particular system of facial recognition system (Amos, Ludwiczuk and Satyanarayanan
2016). The detection of detecting a wide range of few behavioural conditions like
development as well as behavioural disorders is highly require for maintaining ethical
considerations within the organization.
vi) Identification of Anonymous Images: This is yet another important and
significant ethical issue that is required to be considered under every circumstance. The
organizations must not utilize the facial recognition system for the purpose of identification
of anonymous images of one consumer to the next, who could not identify him or her without
even obtaining an affirmative express content.
The ethical issues related to facial recognition system could not easily eradicated by
following few steps. The first and the foremost step is to identify the issues effectively and
efficiently. The regulatory issue or process issue is required to be considered eventually and
hence code of ethics should be compared properly (Happy and Routray 2015). The major
resources of this specific facial recognition system are required to be identified and

9
SECURITY AND PRIVACY ISSUES IN ANALYTICS
information is to be updated effectively. A list of possible actions is made and the positive
and negative consequences.
4. Conclusion
Therefore, it can be concluded that the CTO of the tech start-up has hired a new data
scientist. He has read about Netflix data challenge and wanted to improve the organizational
analytics. A high dimensional sparse dataset is explained in the above report and privacy
issues are identified for that purpose. A contract is being made for building an image
recognition system for law enforcement. The ethical issues are identified for this particular
system.

10
SECURITY AND PRIVACY ISSUES IN ANALYTICS
References
Amos, B., Ludwiczuk, B. and Satyanarayanan, M., 2016. Openface: A general-purpose face
recognition library with mobile applications. CMU School of Computer Science, 6.
Blum, A. and Roth, A., 2013. Fast private data release algorithms for sparse queries.
In Approximation, Randomization, and Combinatorial Optimization. Algorithms and
Techniques (pp. 395-410). Springer, Berlin, Heidelberg.
Cao, N., Wang, C., Li, M., Ren, K. and Lou, W., 2014. Privacy-preserving multi-keyword
ranked search over encrypted cloud data. IEEE Transactions on parallel and distributed
systems, 25(1), pp.222-233.
Chen, R., Fung, B.C., Mohammed, N., Desai, B.C. and Wang, K., 2013. Privacy-preserving
trajectory data publishing by local suppression. Information Sciences, 231, pp.83-97.
Dwork, C. and Roth, A., 2014. The algorithmic foundations of differential
privacy. Foundations and Trends® in Theoretical Computer Science, 9(3–4), pp.211-407.
Greenland, S., Mansournia, M.A. and Altman, D.G., 2016. Sparse data bias: a problem hiding
in plain sight. bmj, 352, p.i1981.
Grinbaum, A. and Groves, C., 2013. What is “responsible” about responsible innovation?
Understanding the ethical issues. Responsible innovation: Managing the responsible
emergence of science and innovation in society, pp.119-142.
Happy, S.L. and Routray, A., 2015. Automatic facial expression recognition using features of
salient facial patches. IEEE transactions on Affective Computing, 6(1), pp.1-12.
Lyu, M., Su, D. and Li, N., 2017. Understanding the sparse vector technique for differential
privacy. Proceedings of the VLDB Endowment, 10(6), pp.637-648.

11
SECURITY AND PRIVACY ISSUES IN ANALYTICS
Martin, K.E., 2015. Ethical issues in the big data industry. MIS Quarterly Executive, 14, p.2.
Mason, R.O., 2017. Four ethical issues of the information age. In Computer Ethics (pp. 41-
48). Routledge.
Mealer, M. and Jones, J., 2014. Methodological and ethical issues related to qualitative
telephone interviews on sensitive topics. Nurse Researcher, 21(4).
Miller, S. and Blackler, J., 2017. Ethical issues in policing. Routledge.
Slade, S. and Prinsloo, P., 2013. Learning analytics: Ethical issues and dilemmas. American
Behavioral Scientist, 57(10), pp.1510-1529.
Uhlerop, C., Slavković, A. and Fienberg, S.E., 2013. Privacy-preserving data sharing for
genome-wide association studies. The Journal of privacy and confidentiality, 5(1), p.137.
Zhu, T., Xiong, P., Li, G. and Zhou, W., 2015. Correlated differential privacy: Hiding
information in non-iid data set. IEEE Transactions on Information Forensics and
Security, 10(2), pp.229-242.