ADVISORY PANEL ON PUBLIC SECTOR INFORMATION Date: Time: Venue: 11 December 2012 10:45am-4:15pm Blue Room -The National Archives Attendees: Chair Members Professor David Rhind, Neil Ackroyd, Expert Member Bob Barr, Expert Member Keith Dugmore, Expert Member David Lammey, Representative Member for Northern Ireland Paul Longley, Expert Member Duncan Macniven, Representative Member for Scotland Hugh Neffendorf, Expert Member Hilary Newiss, Expert Member Shane O’Neill, Expert Member Prabhat Vaze, Expert Member Phillip Webb, Expert Member Nonmembers: 1. Ed Parsons, Geospatial Technologist, Google Annette Dellevoet, Network and Investment Planning team, Sainsburys Dr Harvey Lewis, Research Director, Deloitte Analytics Nigel Clark, Director of Business Development, GB Group Angela Latta, Deputy Director, BIS Data Strategy and Services Team Romina Ahmed, Transparency Team Cabinet Office –(Data.Gov) Alistair Calder, Head of Stakeholder Engagement and Communications Beyond 2011, Office for National Statistics Beth Brook Business and Policy Manager, The National Archives Marcia Jackson, Head of Standards, The National Archives Paul Edwards, Secretariat of APPSI, The National Archives Welcome, introductions and apologies The Chairman welcomed members to the 38th meeting of APPSI. 1.1 The Chairman noted that he had received apologies of absence by members; Michael Jennings, Bill Oates, Patricia Seex, Peter Weinand, and Dean White. 1.2 The Chairman also noted that Carol Tullo Director of Information Policy and Services at The National Archives also had given her apologies through illness. 1.3 The Chairman welcomed Annette Dellevoett, Dr Harvey Lewis, Nigel Clark, and Ed Parsons to the meeting. 1 2. APPSI Seminar Discussion: Private Sector Perspectives on PSI and Open Data 2.1 The Chairman set the seminar discussion around the following questions: Will much business take up happen in the near future? If not, what are the barriers which business sees to it occurring (e.g. costs, documentation) and how might they be overcome? How level is the playing field between public sector users and others? Is there a case for more public data free at the point of use to all sectors? Which are the priority PSI resources? Is the business community content to access very raw data or does it require quality-controlled, refined data? Is continuity of supply an important factor? Should government be collecting different types of data which would be more attractive to business or should government remove itself from the data collection business and leave it to the private sector? To what degree does business need consistent international data sets (e.g. road networks, postcodes / zipcodes)? 2.2 Annette Dellevoett, Network and Investment Planning team, Sainsbury’s Annette began by exploring how a big retail business such as Sainsbury’s utilises data and then offered her personal views on the value of Open Data for the private sector. Annette noted that J Sainsbury plc operates a total of 1083 stores comprising both supermarkets and convenience stores. Over 60% of the UK population lives within 10 minutes of a ’Sainsbury’s store. She explained that the data Sainsbury’s are particularly interested in are those which enable them to forecast sales for new stores, and geospatial data that help them assess acquisition opportunities as well as data that identify competitor trends and activity. She noted that demographic data resource such as Census, ACORN, postcode data and mapping details are particularly critical to their business. In responding to the discussion questions Annette felt that raw data was generally preferable as long as it was clean and well-formatted with good metadata so each field is clearly understandable. In terms of their own business use, a consistent delivery format is essential to facilitate processing the data easily and quickly. For example, CSV, where appropriate, is easy to load into any analysis packages such as GIS tools and statistical packages such as Excel. Annette said that continuity of supply was also crucial for business, as data may be loaded into models or built into in-house analysis tools to be used on an ongoing basis. In some cases this aspect of a dataset may override cost (ACORN versus OAC example). Annette thought that the benefits of Open Data were not only cost savings from ‘free data’ but also the time saved by not dealing with onerous licensing processes to use the data. She argued that open licences would encourage entrepreneurs to develop innovative products from the data and would help to make these products cost-effective to market. Examples include ACORN & Geolytix Postcode Sector Boundaries, both of which were originally built upon PSI. Annette however also argued that continuity of supply was at least as important as cost. In addressing the question of how level is the playing field between public sector users and others,, Annette very much agreed with the premise of calling for more data to be free at the point of use. She noted that many of the datasets covered by the PSMA have been collected using funding provided by millions of citizens and 2.3m businesses through taxation. She explained that by effectively being denied no-cost access to the data the point of use, the customer /user is being charged twice. Annette also personally believed that the collection of new, additional data by government was slightly superfluous 2 and that the government focus should be given to identifying and maximising the value of the data it already holds. 2.3 Ed Parsons Ed Parsons outlined the background to Google’s involvement in the data business. He stressed that their original business model of leasing data from existing providers proved not to be scalable, largely because of the large number of data suppliers involved and the complexity and cost of engaging in the IPR definition process. For this reason, Google had focussed increasingly on gathering its own data. He very much supported the implementation of the UK Open Government Licence which seems to work well but he remained of the view that the no copyright regime (and no licensing) approach used by the US federal government was more effective and certainly easier for business. In fact, he suggested that the level of innovation in the USA was much better due to the public domain approach to PSI with no licence required. A consequence of the Google model being predicated on revenue derived from advertising rather than charging for data use is that they did not know who the end-customers were in many cases: this was true for something around 800,000 of the one million web sites hosting Google Earth. He felt this situation would increase the pressure to develop open data. He explained that forty percent of search terms on Google are geo-related data. In terms of the priority for Open Data Ed therefore argued that the National Address register was a key data set. In response to a question on how to cover the costs of Open address data and how to maintain its quality under such a regime, Ed felt that there was scope to charge for changes to the register with government capitalising the cost over a number of years. 2.4 Dr Harvey Lewis, Research Director, Deloitte Analytics Dr Lewis explained that Delloite had undertaken research on the business perspectives on the use of Open Data for the Open Data Institute. Harvey explained that looking at the supply of and demand for Open Data globally there was no shortage of supply. However the results of their research [see his presentation on the APPSI web site] showed that there is distinct variability of quality and quantity of data provided for reuse in data portals globally. Dr Lewis noted that data.gov.uk for example had a higher number of users, compared with that of France which had a far greater volume of data held on its portal. Harvey explained that business users were far more receptive to the provision of select, high quality data rather than a greater supply of low or medium quality data. Concerns were raised by business responses the research about the defensibility of their product if based upon an open and freely available source. Some of the other key feedback received included that businesses were prepared to pay for data if they knew they would get a better quality of service and consistent and reliable data. Some businesses responded to Deloitte’s research saying that they did not use Open Data at all. A selection of views on Open Data collated from business responses to the Deloitte researchwere: “If you can’t get the data through an API, it’s as good as not being available.” “Businesses don’t want gold-plated data; they want it raw and provided at lowest possible cost.” “Investors are worried about the unpredictability of the environment.” “It takes a long time for information businesses to emerge: this involving stabilising investor decisions, stabilising the data and rendering it into a format that can be used. Only when this is done are you ready to go to market.” “How do you add value to free data? By looking at each sector to identify where the current market proposition is failing.” 3 On barriers to greater use of Open Data, those questioned said these are: • • • • • • • • Lack of awareness of Open Data (and PSI in general) Perceived lack of quality and consistency No guarantees of long-term availability Lack of skills and infrastructure to process data Perceived anti-competitive practices from some public sector bodies Lack of free core reference data Lack of consistency/data standards across different publishers Failure of imagination by larger businesses Dr Lewis argued that for business to engage, raw data is not sufficient: context, quality, consistency, accuracy, and availability all need to be addressed, but only to the extent governed or permitted by the public task of the originating public body. He also felt that lessons need to be learned from successful categories, e.g. transport, and successful practices such as trading funds. Harvey also maintained that, to have greatest economic impact, focus on priority sectors and data categories -- including core reference data - need to be addressed. He concluded that Open Data is the means, not the end in a mixed data ecology and that companies that get the integration right will succeed. 2.5 Nigel Clark GB Group Nigel began by giving some context to the work of GB Group. He explained that they are a company which verifies consumers' identities remotely to help combat ID fraud, money laundering and restrict access to under-age content, purchases and gambling. Nigel said that as part of their operations, they collate data to provide an accurate picture of the UK’s population and property to locate and verify the right –individual [see his presentation on the APPSI web site]. Nigel then went on to discuss the virtues of open data, and admitted limited previous attention to it. He stressed that GB Group did not directly deal with a large amount of Open Data. He said their perception has been that Open Data is very difficult to engage with and often perceived it to be difficult to navigate through to find the data you really want to reuse. Nigel noted that they often found that the information they thought would be useful was provided in either PDF or other unusable formats. He also said there was a perception that a lot of Open Data is incomplete and out of date and unreliable in terms of guaranteed continuity of supply. Nigel believed that overall GB Group was agnostic to the virtues of particular data, stating that they tend to be directed to relevant data by the clients themselves and very little of this is currently what would be defined as Open Data. Nigel strongly argued that the battle of permission in seeking to use Open Data through licences, versus its wide availability to others, meant that the majority of Open Data is not a viable option for GB Group to utilise commercially. Nigel acknowledged though that there is exciting potential in Open Data on matters relating to the following data:: Traffic Times from LAs River Levels and Flow Rates DEFRA claims/payments Divorces Valuation Office property data, Land Registry data number of rooms in dwellings, plus the name of the current registered owner General Register Office Data (Disclosure of Death Registration Information) current and historical– why is this chargeable?) 4 In terms of improving interest in Open Data, Nigel called for: Support to users to appreciate and navigate the data (free!) Central mashing of Open Data Consistency of supply Standard formats (CSV) for relatively raw data Simplified or no licensing As a result of this invitation to consider Open Data, Nigel felt that GB Group would now give it more serious attention. 2.6 APPSI Member observations From the discussions, members observed that there was clear message from these speakers that: The National Address Gazetter should be free at the point of use. Public Sector Information (or some of it) is of real value to the private sector – to both end users, and to resellers (a vital distinction). Its use is currently patchy, both because some data is not yet Open, and also a great lack of awareness. Public Sector Information that isn’t yet Open is underused, not only due to cost, but – in many cases – the complexities of licensing which cause potential users to walk away. APPSI member Hugh Neffendorf added that a clear message from all the speakers was that licensing in many ways can be a bigger perceived barrier than cost, and also that to ensure real engagement with and trust in PSI, business needs to be reassured that there will be continuity of supply. Members noted that a top priority for access to Open Data was the National Address Gazetteer with other property datasets such as VOA, Land Registry and planning applications also being a priority. It was noted that some parts of the private sector want free raw / base data in popular formats (such as csv), but others also welcome the option of resellers offering it in easily used packages. Tracking how data are used and who values it, is a real problem. Many members noted that companies such as Google are more focussed on the platforms that can be created to host the data, rather than analysing what the end user does with it. Many members felt this was a vital issue when seeking to make the business case for investing in PSI (such as the next Census). APPSI members also felt that a clear message given by the speakers is that a stronger emphasis should be made on providing quality data, rather than maximising the quantity of what is available. They noted the view by most of the speakers that a lot of the data available on dgdata.gov is not relevant for business use. Members noted that this view was compatible with APPSI’s focus on prioritising data set availability and the concept of a National Information Framework. ACTION: Secretariat to circulate all presentations to APPSI members 3. Glossary Wiki Exercise developments 3.1 The Chairman welcomed Romina Ahmad to the meeting to discuss the Wiki exercise involving the APPSI glossary. She explained that initial testing of the first pilot had found that the moderating process of the glossary wiki exercise tested in the autumn was cumbersome and that the project needed to be a lot more agile and light touch. 5 3.2 Romina explained that she met APPSI member Bob Barr and the APPSI Secretary to discuss resetting the objectives of the project. 3.2 Romina said that it was agreed that participants of the wiki exercise should be able to provide suggested terms and alternative definitions to those on the list and that these will be submitted for approval then suggestion and discussion online by all participants. Participants should also be able to reference all sources of the glossary and be aware of these sources when feeding back. 3.3 She stated that, whilst it is a consultation to seek validation, it must be recognised that there are already some data terms that are defined in legislation or approved on existing external glossaries. It will thus need to be made clear to participants where the contents of terms and definitions are unlikely to be changed. 3.4 Romina also noted that it is expected that respondents will take part in the consultation online – but that it is acceptable to exploit alternative feedback channels. She explained that the range of potential audiences for the wiki have been identified as including APPSI editorial users as well as Departmental users (in local/national government). 3.5 Romina also set out what the user should expect from the wiki site – these include: Viewing a full list of the Glossary Terms and the ability to select an individual term. Being able to comment generally on each Term and Definition The ability to comment specifically on Terms and Definition and provide proposed amendments to any wording Easy access to provide such feedback. The ability to quote easily from the Terms within the feedback form - citing the whole definition and term, a sentence or an individual word to facilitate making their proposed edits. To be able to see their submission published in the consultation pages (alternative definitions and suggested new terms that will be moderated by APPSI at its discretion) prior to web publication. Alternative Terms that are approved should appear in a separate list with associated comments alongside it. Participants who are feeding back on behalf of an organisation will be encouraged to complete a more extensive web profile in order to aid post consultation analysis. 3.6 APPSI discussed with Romina the moderating requirements for the wiki exercise. It was agreed that the moderators should have administration access to alternative definitions and suggested new terms in order to review all proposals prior to potential web publication. They must also have the ability to decline these ideas and inform participants, as well as posting a holding message on the live site. APPSI also agreed that terms approved by the Moderator should appear in a separate list with associated comments alongside it. There was overall agreement that any moderation by or for APPSIAPPSI should be ‘light touch’ that allows them to review comments periodically. 3.7 Romina said that, in terms of time scales, they would look to develop the new wiki version in January with the intention to test and launch the exercise by Spring 2013. 3.8 The Chairman, Hugh Neffendorf , and Bob Barr agreed that the next step for APPSI was to make the glossary immediately available via the APPSI web pages for general use and comment by email. Following the wiki consultation exercise, the glossary would continue to be available for comment but would probably be fairly stable and only be updated periodically. ACTION: Secretariat to publish APPSI’s glossary work on the APPSI web pages. 6 ACTION: APPSI members Bob Barr and Hugh Neffendorf to liaise with Romina in regard to glossary wiki developments. 4. The National Archives Perspective on current PSI Issues 4.1 Beth Brook provided an update to APPSI on the latest developments in Public Sector Information. Beth noted that since the last APPSI meeting there had been several meetings and compromise texts from the European Council particularly around charging and also cultural institutions. These were based on: European Commission contact with UK Government about its negotiating position and the Open Data agenda. Engagement with HM Treasury on charging and the Ministry of Defence on licensing and scope. Regular engagement with representatives of the cultural sector, with several meetings taking place with the National Museum Directors' Council and Department for Culture Media and Sport, British Library and Arts Council England The outcome of votes by the Industry, Research, and Energy Committee in European Parliament on the amendments put forward. Those agreed centre around data protection (ensuring compliance with the Directive ), formats, open standards and machine-readability, charging and exclusive rights arrangements for digitised cultural resources. Key dates 29/11/2012 – European Parliament: Vote in EP ITRE Committee 17/12/2012 – Trialogues begin between Commission, Presidency and Parliament 20/12/2012 – Telecommunications Council – PSI on agenda as progress report from Cypriot Presidency 10/03/2012 – European Parliament: Vote scheduled at European Parliament Plenary meeting 4.2 4.3 4.4 PSI Directive amendments – Impact Assessment Beth gave an update on the PSI Directive amendments and Impact Assessment (IA). She explained that the appraisal stage for the IA had been completed and sent in late November to Parliamentary Scrutiny Committees and European Affairs Committee in Government for information. Beth noted that several stakeholder workshops had also taken place with the Public Data Group and the NHS regarding the IA. Beth confirmed that the IA will be revised following adoption of the final text of the Directive early in 2013. Information policy Beth noted that the s45 Code of Practice (FOIA datasets section) – Section 45 Code of Practice (dealing with the new datasets provisions in the FOIA sections 11A and 19) was out for consultation online between 21 November 2012 and 10 January 2013. 4.5 Beth also highlighted the Review of PSI (Shakespeare survey) – Stephan Shakespeare appointed as independent chair and Deloittes are carrying out the work, with a commitment to report by January. Several stakeholder surveys had been carried out – including one for PSIHs, one for reusers and a review of existing analyses. 4.6 Beth reported that there was a new Service Level Agreement between the Queen’s Printer for Scotland and the Scottish Parliament, which came into force on 01 October 2012. 4.7 Beth also noted the forthcoming events for Open Data: 7 Transparency Board took place on 21 November 2012. A Licensing Forum was held on the 30 November 2012 for government officials – sessions were presented by BIS from the department for Business Innovation and Skills on the Data Strategy Board and on DCE as well as from the department for Communities and Local Government on data transparency code consultation for local authorities. Open Data Institute launch on 04 December 2012. The first meeting in the UK of OGP’s Ministerial Steering Committee o formally introduced the Independent Reporting Mechanism (IRM), which will monitor progress of participating countries against their transparency action plans, and o Announced the three high-profile figures who will sit on an International Expert Panel to oversee the work of the IRM (former Irish President Mary Robinson, Sudanese-born British entrepreneur and open government campaigner Mo Ibrahim, and Mozambican politician Graça Machel). 4.8 Beth confirmed that the forthcoming events were about to be held: Open Data User Group meeting on 13 December 2012 – Beth noted that Jim Wretham would be updating the group on FOIA s45 CoP (datasets) consultation. Trialogues begin on the 17 December 2012 between European Council/Presidency, and the European Parliament and Commission. Telecommunications Council meeting on 20 December to discuss the Digital Agenda for 2013. 17th meeting of the PSI Group in Europe taking place on 24 January 2013. 5. Minutes and actions of the last meeting 5.1 Because of the necessity to hold the discussion on business views on Open Data in the morning of the meeting, the minutes were taken at this stage in the meeting. Members approved the minutes of the APPSI meeting on 21 September 2012 as an accurate record of the matters discussed. 5.2 Actions from previous meetings: ACTION: Chairman to invite Professor Nigel Shadbolt to a forthcoming APPSI meeting to discuss the Open Data Institute. Status: Complete The March APPSI meeting is to be held at the ODI. ACTION: Bob Barr to continue to refine the glossary document and for it to be reviewed at a future meeting. Status: Ongoing see Item 6. ACTION: Patricia Seex and Phillip Webb to submit those definitions related to data cost and charging for the glossary to Hugh Neffendorf by early October 2012. Status: Completed, version for wiki exercises is ready. ACTION: Peter Weinand to provide commentary on the legal definitions within the APPSI glossary - As above. ACTION: David Rhind to write forward/covering letter for the National Information Framework document. Status: Completed, covering letter drafted and NIF paper now published. ACTION: Secretariat to draft covering submission and forward with covering letter and paper to the minister’s private office. Status: Completed and acknowledgement that the submission had been received and read. 8 ACTION: David Rhind and Hugh Neffendorf and the Secretariat to explore publishing the concept as a journal article. Status: completed David Rhind and Hugh Neffendorf have published the NIF in the Public Leaders Network and Royal Statistical Society web sites. Further publication anticipated. ACTION: APPSI membership to consider how best to engage with the wider PSI community in proselytising the concept. Status: Completed members have shared with the Chairman and Secretariat their intended contacts for the NIF and a consolidated list assembled and used for distribution of alerts to the paper. Further dissemination and activity anticipated. ACTION: David Rhind to reflect upon the Royal Mail issue, and whether a letter to the Shareholder Executive would be appropriate. Status: Completed, letter was sent in October to Stephen Lovegrove. 6. National Information Framework Responses - the next Steps, APPSI Chair David Rhind, and Mr Hugh Neffendorf. 6.1 David informed APPSI that the National Information Framework (NIF) paper had now been published on the Public Leaders Network Guardian website and that he felt progress had been made in disseminating the paper widely amongst officials 6.2 A discussion took place on a general overview of the responses received so far to the NIF document. It was agreed that, while some responses had highlighted areas of the paper that will require further explanation and development of the concept, the overall feel of the responses was that the discussion paper had been quite widely welcomed. 6.3 Hugh Neffendorf very much felt that it was important to maintain momentum of the progress made so far with the paper and to avoid it slipping away. It was agreed that, as the next step, APPSI members should look to consider drafting a series of working papers, to respond in more detail as to the virtues of a NIF and actions required to achieve it. This would be published on the website as supporting information to the main paper. 6.4 It was also agreed that the responses received thus far should be made able via the APPSI website in an appropriate format. ACTION: Bob Barr to propose a suggested initial list of NIF Working Paper topics to provide further detail on the virtues of a NIF. 7. APPSI Planning 7.1 APPSI members discussed the number of meetings for 2013. The Chairman explained that it has been customary to have five meetings a year. He explained this usually means that two meetings are usually close together. Given the active nature of the APPSI membership outside of the meetings, he suggested making the formal meetings quarterly. The APPSI membership agreed to holding quarterly meetings. The Secretariat noted that the proposed meetings dates for 2013 would be 11 March 2013, 18 June 2013, 20 September 2013, 09 December 2013. ACTION: Secretariat to circulate the finalised dates to the APPSI membership. 9 8. APPSI updates - PSI in Scotland, Wales and Northern Ireland 8.1 The Chairman noted that Duncan Macniven provided a written update (See Annex A). 8.2 David Lammey noted that progress had been made with having a representative on the Data Strategy Board and described the outcome. 9. Any Other Business 9.1 The Secretariat confirmed that Neil Ackroyd, Hillary Newiss, Shane O’Neill, Michael Nicholson, Phillip Webb and Patricia Seex had been approved by the Minster for reappointment until 31 December 2015. 9.2 The Chairman noted that this would be Paul Edwards’ last meeting as the APPSI Secretary and thanked him for the support he had given to APPSI. Members associated themselves with the thanks. Meeting Closed: The Chairman noted that the next meeting was due to take place on 11 March 2013 at the Open Data Institute. ANNEX A APPSI MEETING 11 December 2012 SCOTLAND UPDATE The Scottish Government published on 19 September its digital strategy (http://www.scotland.gov.uk/Publications/2012/09/6272). Developed jointly with the Convention of Scottish Local Authorities, the strategy will guide the future actions of the Scottish Government, its agencies and no departmental bodies accountable to Ministers, the NHS, local government, the police and fire services and universities and colleges. The Strategy includes a commitment to: publish as much information as possible concerning the data we hold, and how and when we will make that data available in re-usable form open up access to data created and held by the public sector to make our services more transparent and accountable; open up access to data created and held by the public sector to provide businesses with the opportunities to develop new products and services and therefore grow the economy agree with users and suppliers of ICT systems the most effective common standards to use for publishing and sharing data, to support the linking of data between datasets regardless of source; make data accessible in formats that allow and encourage re-use, with re-use possible under licence terms that are clear, fair, transparent and where possible free." Implementation of the Strategy is overseen by the Digital Public Services National Board and a series of sectoral boards, which also offer a mechanism for future priority-setting. As part of its programme of releasing data in re-useable format, the Scottish Government has published the database of statutory public notices at: http://opendata.tellmescotland.gov.uk. This takes a traditional relational database and exposes it as RDF - with virtually zero cost. 10 The Scottish Government is considering commissioning work on the economic and business value of public sector information to Scotland and the Panel's views on the appropriateness and possible methodological approaches would be helpful. Duncan Macniven 30 November 2012 11