Download presentation
Presentation is loading. Please wait.
Published bySamson Grant Modified over 5 years ago
1
Ethical Implications of using Big Data for Official Statistics
Albrecht Wirthmann, Eurostat, Fernando Reis, Eurostat, 28 June 2018 Session 07
2
Objective Use of Big Data in official statistics
Analysis of ethical issues related to use of big data in official statistics Ethical review Code of Practice Recommendations for using big data in official statistics Ethical guidelines Data linking Secondary data sources Statistical modelling Emerging data market
3
Data sources Telecommunication network data
Sensor data (non-personal, identifiable) Data obtained from internet Financial transaction data Personal health data Electronic reservation systems data Cash register data Telecommunication network data Sensor data that can be related to individual persons or households: Personal communication devices Wearables, e.g. smart watches, heart rate monitors, etc. Smart meters Non-personal sensor data: Road traffic loops Remote sensing data (satellite images, unmanned aerial vehicles (UAV), etc) Data obtained from the internet including: Social media data with or without geolocation information Web-scraped data from company websites, e-commerce websites, job vacancy websites or real estate agencies' websites Query and clickout data from internet searches Financial transaction data (credit cards, debit cards, online payment systems) Personal health data Electronic reservation systems data - e.g. data from flight or hotel booking systems Cash register data, e.g. from supermarkets Crowd sourced data including volunteered geographic data, community picture collections
4
Professional Ethics Declaration on professional ethics
UN Fundamental principles of Official Statistics European Statistics Code of Practice The European Statistics Code of Practice (CoP) defines the principles for the production and dissemination of European official statistics and the institutional environment under which statistical authorities (at national and European level) operate. These principles are in line with the UN Fundamental principles of official statistics and convey the shared professional values of statisticians described in the Declaration on professional ethics, adopted by the International Statistical Institute (ISI) Council in These are respect, professionalism, truthfulness and integrity. Declaration on professional ethics & Eurostat values Respect & Trust Fostering excellence Promoting innovation Service orientation Professional independence institutional environment under which statistical authorities operate Principles for production and dissemination of European statistics
5
Approach Identify issues of the use big data in official statistics
Review the principles and indicators Identify principles of the CoP that give the answers to the ethical questions Consult stake holders of official statistics, to verify with professional values Formulate ethical questions Identify issues of the use big data in official statistics Identify issues of the use big data in official statistics on the basis of ongoing projects and theoretical thinking Formulate ethical questions Consult stake holders of official statistics, to verify with professional values Identify principles of the CoP that give the answers to the ethical questions Review the principles and indicators
6
Characteristics of Big Data
Secondary sources Observing instead of asking Not designed for statistical purposes Held by third parties, mostly private entities Indirect relationship with data objects Emerging data market New analysis methods Machine learning, neural networks Data linking Secondary sources Observing instead of asking Not designed for statistical purposes Held by third parties, mostly private entities Indirect relationship with data objects Emerging data market New analysis methods Machine learning, neural networks Data linking
7
Data acquisition and access
Mandate for data collection Professional independence "Data for statistical purposes may be drawn from all types of sources" Impartiality Public and private interest in using data Mandate for data collection Professional independence "Data for statistical purposes may be drawn from all types of sources" Impartiality Public and private interest in using data
8
Third party data providers
Direct and indirect data collectors Professional Independence of statistical office Emphasis on input quality Possible bias Data manipulation Volatility of data sources Consequences for data provider Reputation Economic consequences Additional burden Reputation of individual and of statistical office Direct and indirect data collectors Professional Independence of statistical office Emphasis on input quality Possible bias Data manipulation Volatility of data sources Consequences for data provider Reputation Economic consequences Additional burden Reputation of individual and of statistical office
9
Privacy & Confidentiality
GDPR for personal data Personal data from non-personal information Threat of disclosure through data accumulation Identification of smaller groups Consent of persons / Information of public GDPR for personal data Personal data from non-personal information Threat of disclosure through data accumulation Identification of smaller groups Consent of persons / Information of public
10
Big Data Analytics Statistical inference from observed behaviour
Statistical concepts Machine learning algorithms "Black box" algorithms Lack of Skills Sourcing Commitment to quality and transparency Adequacy of resources
11
Quality -> Transparency Assessment of input quality
Scientifically sound methodology Data manipulation Inform on output quality Quality dimensions Assessment of input quality Scientifically sound methodology Data manipulation Inform on output quality Quality dimensions Use of big data and data integration -> Transparency Use of big data and data integration -> Transparency
12
Data Market Data as economic asset Data as public good
Algorithms for data analysis Data as public good Right of citizens to be informed Expectation to make profit Exchange of data between private and public actors Data as economic asset Algorithms for data analysis Data as public good Right of citizens to be informed Expectation to make profit Exchange of data between private and public actors
13
Conclusions New principles are not necessary
Amendment or emphasis of certain indicators Similarities with administrative data Private data holders Privacy and confidentiality New aspect related to burden Transparency source data, methods, products Standard setting Role as trusted third party New principles are not necessary Amendment or emphasis of certain indicators Similarities with administrative data Private data holders Privacy and confidentiality New aspect related to burden Transparency source data, methods, products Standard setting Role as trusted third party
14
Ethical Implications of using Big Data for Official Statistics
Thank you for your attention! Albrecht Wirthmann, Eurostat, Fernando Reis, Eurostat, Ethical Review: al.pdf Draft Ethical guidelines:
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.