forum on inn vative data approaches to sdgs. escap_dr. soenke... · • social media data • web...
TRANSCRIPT
Forum on Inn vative Data Approaches to SDGs
31 May - 2 June 2017 Holiday Inn, Songdo, Incheon, Republic of Korea
Innovative Data Approaches for Capturing and Analyzing Data
to Achieve the SDGs
Dr. Soenke Ziesche 31 May 2017
Holiday Inn, Songdo, Incheon, Republic of Korea
Overview of the presentation
• SDGs, targets, indicators and tiers • Big data • Internet of things • Artificial intelligence • Overview of approaches • Challenges • Opportunities • Conclusion
SDGs, targets, indicators and tiers
2030 Agenda for Sustainable Development
Sustainable Development Goals • 17 goals, • 169 targets, • 244 indicators. (232 indicators without duplicates.) Millennium Development Goals 8 goals, 21 targets, 60 indicators.
Indicators
Tier 1 Indicator is conceptually clear and has an internationally established methodology and standards are available. In addition, data are regularly produced by countries for at least 50 per cent of countries […]. Tier 2 Indicator is conceptually clear, has an internationally established methodology and standards are available, but data are not regularly produced by countries. Tier 3 No internationally established methodology or standards are yet available for the indicator, but methodology/standards are being (or will be) developed or tested.
Five indicators with more than one tier assigned. Almost two thirds of the indicators are tier 2 or 3.
Tier 1 2
3
Indicator 82 61 84
Percentage 36% 27% 37%
Big data
Three Vs: • Volume: amount of data • Variety: different types and sources • Velocity: often real-time availability Robert Kirkpatrick, Director of UN Global Pulse: MDG data: Mostly collected and owned by Governments. SDG data: Partly produced passively by people, collected by machines and owned by corporations.
Taxonomy of Big Data:
Exhaust data Passively collected data from people’s use of digital services. Pentland: “It's the little data breadcrumbs that you leave behind you as you move around in the world.” Examples • Mobile phone data • Financial transactions • Online search and access logs • Citizen card • Postal data
Taxonomy of Big Data:
Sensing data Internet of Things and Global Positioning Systems (GPS) Aim: to reduce the information gap between world and internet.
Examples • Satellite and unmanned aerial vehicle imagery • Sensors in cities, transport and homes • Sensors in nature, agriculture and water • Wearable technology (human and animals) • Biometric data
Taxonomy of Big Data:
Digital content Content actively produced by people as well as Governments. Unstructured data, unlike exhaust and sensing data, can include text and multimedia content, e.g. images, videos or audio -> AI analysis Examples • Social media data • Web scraping • Participatory sensing / crowdsourcing • Health records • Radio content
Internet of things
• Sensors: (Often small) objects, which detect changes in its
environment and potentially quantify the extent of the change. • Machine-to-machine communication and AI decision making. • Prevention of negative incidents, e.g. disasters or illnesses,
through early detection.
• Example healthcare: “My car, my airplane, my computer know more about their health status than I do.” Peter Diamandis
Artificial intelligence
“Intelligence measures an agent’s ability to achieve goals in a wide range of environments.” Legg and Hutter (2007)
Artificial general intelligence: - equals human intelligence. - not yet been developed.
Successes were in specialized fields. Examples: Deep Blue (1997), AlphaGo (2016)
Machine and deep learning
Synergies
Overview of big data approaches
Category Source
Exhaust data Mobile phone data Financial transactions Online search and access logs Citizen card Postal data
Sensing data Satellite and unmanned aerial vehicle imagery Sensors in cities, transport and homes Sensors in nature, agriculture and water Wearable technology Biometric data
Digital content Social media data Web scraping Participatory sensing / crowdsourcing Health records Radio content
Exhaust data Mobile phone data
Containing the Ebola Outbreak – the Potential and Challenge of Mobile Network Data SDG targets: 3.3, 3.d Countries: Guinea, Liberia, Sierra Leone
Exhaust data Financial transactions
Scanner data in the Swiss Consumer Price Index: An alternative to price collection in the field SDG indicator: 2.c.1 Country: Switzerland Also: Japan, ROK
Exhaust data Online search and access logs
Big Data: Google Searches Predict Unemployment in Finland SDG target: 8.5 Country: Finland
Exhaust data Citizen card
Oyster card SDG indicator: 11.2.1 SDG target: 11.2 Country: UK Also: China, India
Exhaust data Postal data
Building Proxy Indicators of National Wellbeing with Postal Data 187 countries
Sensing data Satellite and
unmanned aerial vehicle imagery
Japan Aerospace Exploration Agency, Greenhouse gases observing satellite "IBUKI" SDG target: 13.2 Anywhere
Sensing data Sensors in cities,
transport and homes
Sensors to monitor bridges SDG target: 9.1 Country: Sweden
Sensing data Sensors in nature,
agriculture and water
Halt illegal logging in rainforests SDG targets: 15.1, 15.2, 15.b Country: Indonesia
Sensing data Wearable technology
VitalHerd SDG targets: 2.3, 2.4 Anywhere
Sensing data Biometric data
Biometric Cash Assistance SDG targets: 5.a, 8.10 Country: Jordan
Digital content Social media data
Tweeting Supertyphoon Haiyan: Evolving Functions of Twitter during and after a Disaster Event SDG target: 11.5 Country: Philippines
Digital content Web scraping
HealthMap SDG target: 3.3 Anywhere
Digital content Participatory sensing /
crowdsourcing Citizen Feedback Monitoring Program SDG indicators: 16.6.2 SDG targets: 16.6 Country: Pakistan
Digital content Health records
Khushi Baby SDG targets: 3.8, 3.b Country: India
Digital content Radio content
Supporting decision – making through analysis of public radio content Country: Uganda
AI applications towards SDGs without big data
ZenRobotics Recycler SDG target: 12.5 Country: Finland
Challenges
Exhaust data
• Often proxy indicators: Risk of inaccuracy and apophenia.
Sensing data
• Analysis bottleneck, connectivity, interoperability.
Digital content
• Veracity, cleanliness, standards, sustainability.
Overall
• Privacy and human rights
Opportunities
Exhaust data • For human activities that can’t be measured by sensors, e.g. online, financial. Sensing data • Precise and scientific. In the near future many more details of the world and
our daily lives will be measured. Digital content • Innovative tools towards more democracy and maxim “leave no one behind”.
Recommendations
• Governments to explore including big data, internet of things and
artificial intelligence in SDGs National Action Plans.
• Involve private sector, including start-ups, academic research institutions and NGOs to further foster public-private collaboration and knowledge sharing
• Use real time features to reduce the gaps between planning and
evaluation of SDGs activities.
• Pilot appropriate approaches to see whether they may be useful in national contexts.
Group discussion