Has most of government data released openly so far been flytipping, little used and of little value? David Mitton, of ListPoint, will discuss how more value from government data can be realised. Using the national police database as a case study, he will examine how this data was made more widely usable and how IT implementation costs were reduced by 20%, just by surfacing and mapping the data standards.
2. Reference data (Code Lists)
County Codes County Codes
(Local Gov Standard) (Central Gov Standard)
Kent KT101 Essex ESX
Essex EX101 Herts HFD
Hertfordshire HT101 Beds BD1
Bedfordshire BD101 Cambridgeshire CB
Cambridgeshire CB101 Kent KNT
Mapping competing codes together creates a data translation layer
Federated data is made interoperable without redefining data standards
3. Fly Tipping
Picture: Kevin Rothwell, Creative Commons
• Wasteful
• Drain on resources
• Costly
5. Recycling
• Social and economic benefits
Reduction in pollution
Creation of 4 :1 extra jobs versus traditional incineration
Eases demand for natural resources
6. Fly Tipping
Homeless Temp Leasehold
Local Authority B&B Hostels Admin % Admin
Spending Accom Dwellings
Barnsley Borough Council £396,000 £396,000
100%
Doncaster Metropolitan Borough Council £489,000 £489,000
100%
Knowsley Metropolitan Borough £547,000 £547,000
100%
Newcastle upon Tyne City Council £1,239,000 £1,239,000
100%
Oldham Metropolitan Borough Council £395,000 £395,000
100%
Sunderland City Council £876,000 £876,000
100%
North Tyneside Council £248,000 £0
0%
St Helens Borough Council £399,000 £0
0%
"from OpenlyLocal"
• Wasteful
Resource to publish and high cost to interrogate
Where to start?
7. Dumping
Homeless Temp Leasehold
Local Authority B&B Hostels Admin % Admin
Spending Accom Dwellings
Bradford Metropolitan Council £1,459,000 £472,000 £987,000
68%
South Tyneside Metropolitan Borough Council £344,000 £129,000 £215,000
63%
Sefton Metropolitan Borough Council £257,000 £33,000 £65,000 £160,000
62%
Sheffield Council £5,141,000 £2,109,000 £0 £3,032,000
59%
Coventry City Council £676,000 £345,000 £331,000
49%
Salford City Council £1,695,000 £239,000 £507,000 £137,000 £812,000
48%
Leeds City Council £9,489,000 £2,937,000 £527,000 £1,723,000 £4,302,000
45%
Wigan Metropolitan Borough £820,000 £69,000 £399,000 £352,000
43%
Rotherham Metropolitan Borough Council £785,000 £25,000 £501,000 £140,000 £119,000
15%
"from OpenlyLocal"
8. Organising
• The standards against which they report
Bed & Breakfast
Hotels
Administration
Temporary Accommodation
• The reference data to classify it
Ability to map reference data
A means to make sense of diverse data sets
• Outcomes
Improved quality and reusability of data
• Data.gov.uk
9. How important is being
organised?
• The Soham Murders
• Holly Wells & Jessica Chapman
• Data was not interoperable
• Information existed to prevent this tragedy
10. The challenge of interoperability
M = Male, F = Female, U = Unkown….or does it?
• Try to enforce standards and you get resistance
• Solution is to manage diversity, not try to impose uniformity
• Good standards emerge by embracing all standards and
collaborating
11. The Police National Database
PND Outcomes
National View of Criminal Police National Dbase
Intelligence 300 code list standards
£2,500,000 savings 4. Translation layer (XML)
on data integration CRASH
PentiP
5. Quality Control 6. Re-use of standards HOLMES 3
HODH
MoJ
Alerts Web Services
3. Map to PND Police.uk
2. Application Context
1. Validate Local Codes
300 ICT systems and 43 police forces = 12000 competing code lists
Gender Classifications Male = M, Male = M1, Male = 001, Male = M101
14. The value of one data standard
Vehicle Make Model Code List
Government Private Sector
Police Insurance Companies
SAFETY VALUE QUALITY
VOSA
CRASH statistics Trade – e.g. RAC, Tracker
Aftermarket parts MOT Failure Codes
DVA NI Rental & Leasing
VRM Lookup companies
Department for Transport 40 x Motor Manufacturers
1000’s Parts suppliers
15. It seems obvious. So why is data
‘dumped’ and savings missed?
• Only 35% of UK civil servants know that using data standards
can reduce IT implementation costs
• Only 40% know that data standards make data interoperable
• 78% do not know what benefits will flow from the open data
agenda
Source: Listpoint/ Dods civil service research into understanding of open data agenda and benefits –
December 2012. More information at www.listpoint.co.uk
16. What can the OD community
do to give data more value?
• Government Open Data is not Open Data
without the reference data sat alongside it
• Champion the cause
Raise awareness of the role classifications play
• Re-use the tools available
• Outcomes
Reduce costs, improve services, innovation
Greater value in ‘Open Data’
17. £655 Million in savings if code
lists are organising and reused
GOVERNMENT SPEND ON IT PROJECTS (RENEWING)
2013 2014 2015 2016
Department ICT Value No. ICT Value No. ICT Value No. ICT Value No.
Attorney General’s Office
Cabinet Office
Department for Business, Innovation and Skills £240,000 1 £260,000,000 3 £125,000,000 2
Department for Communities and Local Government £5,000,000 1 £800,000,000 2
Department for Culture, Media and Sport £14,000,000 1
Department for Education £15,000,000 2 £90,000,000 2 £25,000,000 1
Department for Environment, Food and Rural Affairs £15,000,000 1 £11,500,000 1 £500,000,000 1
Department for Health £106,700,000 4 £48,200,000 1
Department for Transport £153,700,000 3 £946,200,000 5 £82,000,000 1
Department for Work and Pensions £135,000,000 1 £687,000,000 3 £100,000,000 1 £300,000,000 1
Department of Health £149,000,000 2 £6,680,000,000 12
Foreign and Commonwealth Office £82,000,000 1
Greater London Authority £270,000,000 1 £38,000,000 2 £12,000,000 1
HM Revenue and Customs £1,610,000,000 2 £56,000,000 1
HM Treasury £100,154,000 3 £80,000,000 2
Home Office £122,000,000 1 £30,500,000 1 £10,000,000 2 £1,328,100,000 4
Local Authority £687,396,347 30 £351,397,244 22 £364,225,000 33 £896,200,000 14
Ministry of Defence £23,500,000 3 £173,000,000 2 £6,895,000,000 6 £80,000,000 1
Ministry of Justice £501,410,000 3 £62,000,000 2
Police £4,800,000 1 £0 1 £8,430,000 4 £600,000,000 2
Transport for London £10,500,000 1
Welsh Assembly £274,000,000 2
Other £9,000,000 3 £12,000,000 1
Grand Total £3,869,246,347 63 £8,744,551,244 56 £9,575,455,000 64 £4,052,500,000 30
Average spend on data integration = 12.5% of total £483,655,793 £1,093,068,906 £1,242,460,000 £506,562,500
Savings by adopting Listpoint (as per PND) £96,731,159 £218,613,781 £248,492,000 £101,312,500
Savings over 4 Years on ICT projects £665,149,440
18. Appendix
• Listpoint Background
• Listpoint Features
• Reference data mapping
19. Listpoint Background
• 16,000 Government reference data sets published
• Classified as a re-usable asset on the ICT ASK register
ICT Asset and Services Knowledgebase
Part of the ERG process (Cabinet office review of planned £5m+
government projects – to ensure re-use and not re-invention)
• A member of the European Commission CESAR programme
Community of European Semantic Asset Repositories
Largest contributor of semantic ‘assets’
• Award winning application
Emergency Services Awards 2011
• A free to use platform to find, load, validate and collaborate
around reference data standards
20. Listpoint in practice
• Code List quality
Browser-based editing/publishing and integrated validation rules
• Context management
Group relevant code lists together
Recognise in which ‘context’ a data standard is being used
• Context Mapping
Listpoint approach to semantic interoperability
Join competing standards together with auto mapping)
• Quality maintenance
Alerts (e-mail and web services) when standards are updated
• Translation
Output includes XML and SKOS – XML mappings used to translate data
across multiple applications for efficient data integration
Editor's Notes
It is a big ask to get departments to align the standards to which they report but it does exist – take central government reporting requirements of local governmentHowever – the reference data existsData.gov.uk has over 9000 data sets on it, but if you search for ‘reference data standard’ – you get Zero. Search for code list – and you get 1200 returns but very few are the actual reference data that sits alongside the data There are over 9000 datasets tipped into data.gov.