4. 4
1. Aging, cities, demography
2. Energy, climate change and the natural
environment
3. The economy, the City, the world in trade
4. Manufacturing, materials science, innovation
5. Data, infrastructure, identity
6. Principles of government decision-making:
risk, resilience, contingency
Themes of GO-Science work
All activities need to be underpinned by evidence,
analysis and data
Science, evidence and data in government, 3rd
September 2013
5. 5
Bringing a
multidisciplinary
approach to the
challenges facing
government
Dave Ramsden
1,350
Economists
Tony O’Connor
450 Operational
Researchers
Jil Matheson
1,600
Statisticians
Jenny Dibden
950 Social
Researchers
Mark Walport
12,000 Scientists
& Engineers
Science, evidence and data in government, 3rd
September 2013
6. 6 Science, evidence and data in government, 3rd
September 2013
A network of Chief Scientific Advisers
Prof
Chris Whitty
DFID
Prof Dame
Sally Davies
DH
Prof Robin
Grimes FCO
Carole
Willis
DfE
Prof John
Perkins
BIS
Prof David
Mackay
DECC
Prof Rod
Smith
DfT
Dr Bill
Gunnyeon
DWP
Dr James
Richardson
HMT
Prof Bernard
Silverman
Home Office
Prof Peter
Freer Smith
Forestry C.
Dr Andrew
Wadge
FSA
Rebecca
Endean
MoJ
Dr David Bench
HSE
Prof Julia
Slingo
Met Office
Prof Vernon
Gibson
MOD
Prof Ian
Boyd
DEFRA
Prof Bernadette
Hannigan
NI Interim CSA
Prof Muffy
Calder
Scotland
7. • Data
• Information
• Knowledge
• Application
7
From data to knowledge to society
Science, evidence and data in government, 3rd
September 2013
GrowthResilience
Policy &
service
delivery
Societal Benefit
9. • Defining & refining the
question
• Data collection
• Analysis
• Visualisation
• Inference
Information
9
From data to information using statistics
Science, evidence and data in government, 3rd
September 2013
Application of
Statistics at all
levels
Communication
11. 11 Science, evidence and data in government, 3rd
September 2013
• One of the oldest
continuous agronomic
experiments in the world
– started in 1843
Old ways of collecting data
Herbicides and fertilisers Herbicides but no fertilisers No herbicides
Rothamsted’s Broadbalk
Winter Wheat Experiment
12. • Data is being collected in different ways
• Users submit photos and locations of sightings to a
team who refer them on to the Forestry Commission,
which is leading efforts to stop the disease's spread
with the Department for Environment, Food and Rural
Affairs (Defra)
12
Chalara spread: 1992-2012
New ways of collecting data
Science, evidence and data in government, 3rd
September 2013
13. 13
• Developed by Boston's Mayor's Office of
New Urban Mechanics
Built environment
• Street Bump uses crowd
sourcing to improve
neighbourhood streets by
collecting road condition data
while they drive
• Utilises two of the phone’s
sensors, its accelerometer and
GPS. The sensors detect
“bumps” that the City maps
Science, evidence and data in government, 3rd
September 2013
14. 14
Long-term climate data
Science, evidence and data in government, 3rd
September 2013
mean
(1961-1990)
= 0
Source: Met Office Hadley Centre (2013)
Global historical surface temperature anomalies
16. Geoposition data
16 Science, evidence and data in government, 3rd
September 2013
Source: Sir
Stuart Peach,
Cambridge
Conference
2013
Tactical level analysis: Ditch & Flow in
Helmand
Tribal Distributions
Population density
Population movements
Terrain Elevation Data
Topographic Data
Combined Operating Picture
Oceanographic Data
Hydrographic Data
Meteorological Data
Aeronautical Data
Geodetic Data
CrawlStalkWalk
17. Analysing big data
• Total Scottish Population 5.2M
• People with diabetes : 251,132 (4.9%)
• People with Type 1 DM : ~27,000
(0.5%)
• All patients nationally are registered
onto a single register; the SCI-DC
register
• SCI-DC used in all 38 hospitals
• Nightly capture of data from all 1043
primary care practices across Scotland
Courtesy of Andrew Morris
17 Science, evidence and data in government, 3rd
September 2013
Harnessing ICT: A national diabetes system for Scotland
18. 18
PercentageofPatients
Data recorded within the previous 15 months
http://www.diabetesinscotland.org.uk/Publications/SDS%202010.pdf
Courtesy of Andrew Morris
Scottish Diabetes Survey – over 90% capture of key variables since 2007
Recording of Key Biomedical Markers
Science, evidence and data in government, 3rd
September 2013
Accountability of services
19. Diabetes Care 2008Diabetic Medicine 2009
Courtesy of Andrew Morris
Improved clinical outcomes
19 Science, evidence and data in government, 3rd
September 2013
20. 20 Science, evidence and data in government, 3rd
September 2013
Demography
Future of Demographic Change project:
• A growing population
• An ageing population
• Local variation at all scales
• Increasing diversity
• Demographic changes will have an
impact
Graphic by The Guardian, 2010. Source data: ONS, Population by Local Authority, 2010
UK population change 1999-2009 by local authority
21. Polar area diagram by Florence Nightingale illustrating causes of mortality during the Crimean War (1857)
21 Science, evidence and data in government, 3rd
September 2013
Turning data into information
22. The UK may be considered to be a small emitter
compared to some other countries
22
Source: The Carbon Map
Tonnes
CO2 per
capita
Science, evidence and data in government, 3rd
September 2013
Visualising emissions data
23. But looking at historical emissions tells a different story
23
Source: The Carbon Map
Tonnes
CO2 per
capita
Science, evidence and data in government, 3rd
September 2013
Visualising emissions data
25. 1.Communication and misuse
2. Privacy
3. Skills
25 Science, evidence and data in government, 3rd
September 2013
Challenges
AndyHedges / CC BY-SA 3.0
26. 26 Science, evidence and data in government, 3rd
September 2013
• Need for us all to be clear in communicating
• Especially where:-
- the science or statistics are complicated
- there are uncertainties
- big press interest
- emergency situation
• Direct language and clear diagrams – but without
losing the nuances
Communication: uncertainty and
getting our message across
27. 27 Science, evidence and data in government, 3rd
September 2013
• ‘Cone of uncertainty’ diagram – line shows most
likely path of hurricane and the uncertainty
displayed by the white region indicating a 2 in 3
chance of containing the path
• But the most likely path provides false sense of
security to people living away from the black line
• For Hurricane Irene in 2011 a ‘spaghetti plot’ of
possible paths were shown on TV, each path arising
from a different forecasting model
• Communicates uncertainty of the path better than the
‘cone of uncertainty’
Communication: uncertainty and
getting our message across
28. High emissions scenarios project warming of
around 4°C by the end of the century, strong
mitigation scenarios limit this to 2°C
28
• Even in the strong
mitigation scenario we
are still committed to
further climate change
for the next 2 to 3
decades, due to inertia
in the climate system
• How much the climate
warms after mid-
century depends on
how long global action
on mitigation is delayed
Science, evidence and data in government, 3rd
September 2013
Uncertainty of predictions
29. Risk vs Hazard
Radiation
Dosage
Chart
Sources:
InformationisBeautiful.net
BBC, Guardian Datablog, Mayo Clinic
data: bit.ly/radiationchart
Micro-Sievert (µSv)
0.1 µSv
1.0 µSv
10 µSv
100 µSv
1000 µSv
0.1 – Eating a banana
0.4 – Natural radiation in human body
40 – Flight from New York to LA
70 – Living in a stone, brick or concrete building
for a year
100 – Chest X-ray
250 – Release limit for a nuclear power plant for a
year
400 – Yearly dose per person from food per year
1000 – EPA yearly limit on radiation exposure
to member of public
1.0 – Using CRT monitor for a year
3.5 – Extra dose from one day in average town
near the Fukushima plant
5.0 – Dental X-ray
10 – Background dose received by average
person on an average day
29 Science, evidence and data in government, 3rd
September 2013
30. 30 Science, evidence and data in government, 3rd
September 2013
Source: http://understandinguncertainty.org/micromorts
Risk communication – Micromort
• Comparing
small lethal
risks (acute
risks) difficult
• Micromort: 1
in a million
chance of
death
• Translates
risk into easily
comparable
units
Walk
Cycle
Motor
bike
Car
Micromorts per 100 miles travelled
31. 31 Science, evidence and data in government, 3rd
September 2013
• There are many “publics”
• Public values play a key part in how people respond
• Predictable “fright factors” – feelings of lack of control
e.g. multinational companies’ monopoly of a new
technology
• Need to build “trust” (impartiality/credibility of the
commentator)
• Acknowledge and respect the “feelings” of those
responding
Communication: understand
your audience
32. Public values on energy
Summary of core public values on energy system change
32
Source: UKERC Transforming the UK Energy System: Public Values,
Attitudes and Acceptability
• 73% of respondents agree Britain should reduce
energy use
• 79% of respondents say UK should reduce use of
fossil fuels. Predominant reason was unsustainable
nature of fossil fuels and environmental harm
• 83% of respondents are fairly or very concerned that in
the next 10-20 years electricity and gas will become
unaffordable for them
Science, evidence and data in government, 3rd
September 2013
• To understand public preferences need to understand
public values
• Public preferences change with context, but values
and principles guide decisions and engagement
33. 33 Science, evidence and data in government, 3rd
September 2013
Misuse of statistics: Visualising data not
propaganda
• Figures are factually correct
• But infographic is misleading
• From the North South Wales
Ministry of Health, March 2013
Source: Guardian Data Blog
34. 34 Science, evidence and data in government, 3rd
September 2013
Misuse of statistics: safeguards
• The UK Statistics Authority – Whitehall’s statistics watchdog
• Chairman Andrew Dilnot can comment on perceived misuse
of official statistics
• Groups such as Straight Statistics, Sense About Science and
online blogs (e.g. Ben Goldacre’s Bad Science) now work to
highlight misuse of evidence
• Promoting a culture of open access and providing links to
referenced studies will allow for people to scrutinise the statistics
• Improving the accessibility of the ONS website
• Improve mathematics and statistics in education (as suggested
in the 2012 CST letter on STEM Education)
35. 35
• Personal information is individual &
precious to each one of us – it’s vital that we
treat it properly
• A balancing act…between the right to
privacy and the necessity to hold and share
data
• A framework is needed:
- to protect individuals
- build & maintain confidence
- facilitate research
Privacy
Science, evidence and data in government, 3rd
September 2013
Wellcome / BY-NC-ND 2.0
36. 36 Science, evidence and data in government, 3rd
September 2013
Skills pipeline
Source: DfE
Mathematics teacher recruitment shortfall/surplus
37. 37
CST work: The Age of Algorithms
Schools Universities Jobs
• Recruit more teachers trained
in computer science
• Embrace Continuing
Professional Development
(CPD)
• Incorporate logic, algorithms
and statistics into post-16
mathematics courses
• Government should sponsor new Computer Literacy Campaign highlighting learning
opportunities for individuals, schools and businesses
• Cooperation between
computer science
departments and private
sector to develop
multidisciplinary courses
for practical application of
data science
• Create a National
Centre to promote
advanced research
and translational work
in algorithms and
application of data
science
Science, evidence and data in government, 3rd
September 2013
38. 38
We all need to work together
Science, evidence and data in government, 3rd
September 2013
Universities Industry
39. 39
How can we help each other
Science, evidence and data in government, 3rd
September 2013
Central Intelligence Agency
40. Every effort has been made to trace copyright holders and to obtain their permission for the use of copyright material.
We apologise for any errors or omissions in the included attributions and would be grateful if notified of any corrections
that should be incorporated in future versions of this slide set. We can be contacted through enquiries@bis.gsi.gov.uk .
Science, evidence and data in government
Presentation made to the Royal Statistical Society annual conference,
September 2013
Sir Mark Walport, Chief Scientific Adviser to HM Government
Editor's Notes
In January 1941, the Central Statistical Office (CSO) was established to co-ordinate all statistics brought before the War Cabinet. Harry Campion became its first director. He remained its director – not counting some secondments to the United Nations – until his retirement in 1967 The CSO became established as a permanent feature of government focusing on the development of national income accounts.. After WWII, the United Nations Economic and Social Council (ECOSOC) invited a number of prominent national statisticians, including Harry Campion, to form a 'Nuclear’ Statistical Commission. It made recommendations for the terms of a permanent commission on statistics; the organisation of a statistical organisation; the existing statistical activities conducted by the League of Nations ; the relationship between the statistical activities of the UN and specialized agencies . He later became its Chairman for 1951 and 1953. After the war, Campion helped to organise the creation of the United Nations Statistical Office . He was President of the Royal Statistical Society from 1957-1959, and in 1996 he left a bequest. These funds have enabled both the Campion Fellowship – awarded to Fellows of the Society for a project towards the development, use or exposition of statistics on the economic or social well-being of the population – and the RSS Historical Fund , which supports the conservation and restoration of the Society archives and book collection, to be established. Other posts he held included President of the International Statistical Institute and Vice-President of the Manchester Statistical Society .
The Top Priorities for GO-Science.
Winter wheat has been sown and harvested on all or part of the field every year since then Original aim was to test the effects of various combinations of inorganic fertilizers (supplying the elements Nitrogen, Phosphorous, Potassium, Sodium and Magnesium) and different organic manures on the yield of winter wheat. A control strip has received no fertilizer or organic manures since 1843 Originally the weeds were controlled by hand weeding but later by periodically bare-fallowing and cultivating different parts of the field in different years. From the mid-1950s, herbicides have been used. Two major modifications were made from 1968: (1) introduction of modern, short-strawed cultivars. (2) crops other than wheat being grown on the experiment, so that yields of wheat grown continuously could be compared to those of wheat grown in rotation. Tile drains under each treatment strip are used these to collect and measure the nutrients in the water that leached through the soil into the drains Untreated wheat stays relatively constant with a yield of 1 tonne per hectare The improved cultivars, cultivations and control of pests, diseases and weeds have demonstrated that with the right combination of treatments the wheat yield can be increased to 9 tonnes per hectare
Ash dieback, caused by the fungus Chalara fraxinea, was found in the UK in October outside of plantations and nurseries in East Anglia, raising fears of a repeat of Dutch elm disease which killed 25 million mature elms in the 1970s and 80s. In an attempt to map and help prevent the spread of the disease across the country, a team of developers and academics worked through the weekend to create an app that smartphone owners can use to report suspected cases of infection. Infected ash trees are recognisable by lesions on their bark, dieback of leaves at the tree's crown, and leaves turning brown – though experts say the arrival of autumn makes the latter harder to accurately spot. zThe AshTag app for IOS and Android devices allows users to submit photos and locations of sightings to a team who will refer them on to the Forestry Commission, which is leading efforts to stop the disease's spread with the Department for Environment, Food and Rural Affairs (Defra).
HadCRUT4 is a gridded dataset of global historical surface temperature anomalies relative to a 1961-1990 reference period. Data are available for each month since January 1850, on a 5 degree grid. The dataset is a collaborative product of the Met Office Hadley Centre and the Climatic Research Unit (CRU) at the University of East Anglia. Brief description of the data The gridded data are a blend of the CRUTEM4 land-surface air temperature dataset and the HadSST3 sea-surface temperature (SST) dataset. The dataset is presented as an ensemble of 100 dataset realisations that sample the distribution of uncertainty in the global temperature record given current understanding of non-climatic factors affecting near-surface temperature observations. This ensemble approach allows characterisation of spatially and temporally correlated uncertainty structure in the gridded data, for example arising from uncertainties in methods used to account for changes in SST measurement practices, homogenisation of land station records and the potential impacts of urbanisation.
Geospatial information: Provides focused briefing support Catalogue, organise and understand existing data Warn of environmental and other risks Provide feedback Improve government effectiveness and efficiency Opportunities Existing partnerships can be enhanced Collaboration and Burden Sharing. Integration to add value Dissemination in time is critical Integration of Geospatial Foundation Intent to integrate Geo Foundation data, products and services to provide a more complete picture of the operating environment. Tactical level product. Original requirement and specification developed for JNAC (Joint Narcotics Analysis Centre at OWOB) for poppy eradication. Information was required on all ditches (main & field / widths & depths) in the Nad ‘Ali and Marjah (west of LKG). 2 versions produced (Analysis-classified & Plans-unclass). Stereo source used for capture to the nearest meter. This allowed assumptions to be made of those involved in opium trafficking and IED component smuggling in using the ditches as routes in and out of the area by concealed walking, stalking or crawling, and their travel times. Includes tree lines. Time / distance can be estimated e.g. walking at 4 kmph; stalking at 2.5 kmph and crawling at 1 kmph. This product has been broadened for Theatre use and used to plan alternative routes in the intensely irrigated areas using gap crossing bridging equipment, particularly by quad bikes. Widths from 1 to 32m. Depths from 0.5m to >1.5m. Each area contains approx 6,000 ditches & 4,000 culverts. Areas cover are Nad ‘Ali, Marjah N & S, CAT, Babaji and Bolan Desert (west of LKG). A further RFI has been received (May 10) from HQTFH to complete Ditch Analysis for the full TFH AO. Completed Oct 10.
Future of Demographic Change project: How might global demographic trends impact on the UK? The UK National Population Projections indicate: Continued growth for the whole UK from 62.7 million people at the end of 2010 to 73.2 million by the end of 2035 The age structure will become gradually older the median age rising from 39.7 years in 2010 to 42.2 years in 2035 Local variation at all scales Patterns of demographic change at ‘sub-national’ level continue to be highly variable. Increasing diversity Net inward-migration projected beyond 2060 means the UK will continue to become more ethnically diverse UK national issues: Demographic changes will have an impact on all areas of public expenditure Recent birth-spike will have knock-on effects through the system Key policy areas will all be effected by demographic change, but policy decisions made will also drive future change in demographic patterns This interaction of policy and wider trends limits the accuracy of future forecasts ONS - “Beyond 2011” working to develop the options for future population and socio-demographic data collection. Data reliability becomes more difficult to ensure affordably at smaller scales – some areas are subject to greater ‘churn’ than others. Data definitions need to be better specified to improve consistency of primary data. Greater strategic use of UK patterns of demographic change can provide a more robust analytical baseline to inform current and future policy needs. International demographic data can be better used to further UK strategic interests. Use of systems modelling could enhance understanding of demographic impacts and reduce unintended consequences in policy decision making – this concept to be tested as part of the project.
This map shows country sizes according to historical CO2 emissions (1850–2007) Historical (or 'cumulative') emissions are relevant because CO2 is a long-lived atmospheric gas. Europe and the US dominate historical emissions, having released around half the CO2 ever emitted. For historical emissions Great Britain ranks 5th out of 187 countries (~6 % of global historical emissions)
Representative Concentration Pathway or RCP
Yet it is important to keep hazards in perspective…how does Fukushima compare?
David Spiegelhalter developed this graph showing micromorts for type of travel
Need to look beyond public preferences because these are likely to change depending on context, particularly considering highly unfamiliar issues where perceptions and preferences are not yet fully formed. Examine the values and principles that people draw on to guide decisions and engagement. Efficiency and not wasting – in sum, being more efficient (doing more with less) and minimising waste and overall energy usage is almost universally seen as positive. Protection of the environment and nature – in sum, being environmentally conscious and respectful of nature through minimising intrusive and destructive processes. Ensuring security through reliability, affordability, availability and safety – in sum, making sure the energy system is safe, reliable and accessible to citizens, both in terms of personal affordability and national availability. Autonomy and power – in sum, being mindful of the importance of autonomy and freedom both at national and personal levels. Social justice and fairness – in sum, developing energy systems in ways that are open, transparent and fair and attentive to the effects on people’s abilities to lead healthy lives. Improvement and quality – in sum, thinking in terms of long term trajectories, ensuring changes represent improvement and considering their implications for quality of life. Also, tensions exist between values (how people think things should be), and world views (experiences or perceptions of how things currently are). Values also interconnect with people’s life experiences and social commitments (e.g. their relationships with others, their form of work).
This infographic from New South Wales shows the increase in the number of nurses - which is factually correct. The only problem is that you get a very different sense of scale from studying the image than you do if you just glance at it. Four stick people represent 43,000 nurses - so why are 28 more stick people used to represent an increase of just 3,000 nurses? That's a 700% infographic explosion to show a 7% increase. We can only assume someone leant on the paste key.
Privacy The use of datasets containing personal information is vital for a lot of research in the medical and social sciences, but poses considerable challenges for information governance because of the potential to compromise individual privacy. Citizens have a legitimate interest in safeguarding their privacy by avoiding personal data being used to exploit, stigmatise or discriminate against them or to infringe on their personal autonomy (see box 3.4).159 The legal framework for the “right to respect for private and family life” is based on article 8 of the European Convention on Human Rights (ECHR)160 for member states of the Council of Europe. Some aspects of privacy rights are codified by the EU Data Protection Directive (95/46/EC) and implemented in the UK by the Data Protection Act 1998 (DPA). There is a live issue with the EU Data Protection Regulation, specifically that the research community are very concerned that amendments proposed by the rapporteur of the LIBE committee will prevent or severely impair scientific research studies using personal data. BIS and DH are working with MoJ (who are in the lead on negotiations) to address. Series of public perception flashpoints e.g. 2 CDS 25m items of child benefit data (2007), Google Maps, Street View.
Hopefully this presentation has demonstrated the obvious importance of analysis and presenting data well The RSS brings together a strong group of talented statisticians Tell me: What can you do for me and what can I do for you? Data without analysis and effective communication is useless We need each other