Addressing
|
Note: When you edit this page, you agree to release your contribution into the public domain. If you don't want this or can't do this because of license restrictions, please don't edit. |
|
TODO: what's the purpose of this topic, goals, etc.
Contents |
Attendees
Rapporteur: Prof. Rob Kitchin, NUI Maynooth Group:
- Jim Cuddy (Enterprise Ireland),
- Ciaran Dynes (Talend),
- Annette Farell (IRLOGI),
- Justin Gleeson (NUIM),
- Enda Holland (PER)
Data quality and standards
Spatial data quality in Ireland is widely acknowledged as being very poor. In the absence of postcodes, address information supplied to public agencies by citizens and companies is highly variable in accuracy, quality and format, especially for rural areas where there are no street names or numbers and where there might be multiple instances of townlands sharing the same address in each county. As yet, there are no universal data standards operating for addressing in Ireland. As such, the fields of address data collected by agencies can vary between organisations. What this means is that it can be difficult and time consuming to georeference data, even using a product like Geodirectory, and even when it is undertaken there are issues of accuracy and reliability. Where public administration data are available, they are typically released at very coarse scales, usually counties or similarly sized units such as garda districts or health board areas, and at best electoral districts (EDs). The result is that it very difficult to map individual address points or conduct localized analysis of public data. One significant recent advancement is the new finer grained administrative geography recently rolled out by Ordnance Survey Ireland (OSI). The new Small Areas relate to areas of c.100-125 households. They have been constructed in such a way that they fit neatly inside every other administrative boundary and also do not cross significant natural and people-made boundaries such as rivers and major roads. The 2011 census will be released at the Small Area level, but they are presently little used beyond the CSO because without postcodes or the routine generation of georeferenced data it is time consuming or expensive to assign data to Small Areas. Recommendations: (1) the introduction of data standards with respect to addressing be adopted by all public agencies; (2) the creation of a free to use lookup table to georeference addresses; (3) the routine outputting of public administration at the finest spatial granularity possible (preferably Small Areas or amalgams thereof) taking into account confidentiality issues.
Accessing data
There is a massive amount of data residing in public and private bodies that is presently under-used and under-analyzed. A key objective of the opendata initiative is to liberate these data so that they can be used by citizens, companies and the state to derive benefits. Accessing georeferenced and cartographic data in Ireland is limited by three factors: commercial ownership, copyright and intellectual property rights, and data licenses. In the absence of postcodes, comprehensive fine-scale (individual address point) georeferencing is only available through Geodirectory (a commercial company owned by An Post and Ordnance Survey), Loc8 (a private postcodes company), and specialist address matching companies. There is no free address look-up table as with other Western countries. Accessing Geodirectory is prohibitively expensive for citizens, small companies and small public bodies. Moreover, there are data licence restrictions on how the data and derived data can be displayed, shared, used and published that limits potential application. To date, there has not been a crowdsourced movement to create a free, opensource address look-up table. The situation with respect to cartographic data is a little better. Openstreetmap, Google Maps, Bing Maps and others make mapping data widely and freely available through APIs, but is limited in its scope (principally road data) and focus (locating places), and in the case of Openstreetmap its areal coverage. Ordnance Survey generates and maintains a greater variety of data, including administrative boundaries, land parcels, terrain models and land-use. In general terms, OSI generates the majority of its revenue from a minority of its data, and yet the most useful data in many ways for the opendata initiative is the locational and administrative boundary data that generates little of its revenue but is required for outputting public data held by the CSO and other agencies. At present, the use of all OSI data is limited by data licensing and copyright protection meaning that it cannot be used for opendata projects without the agreement with OSI. Recommendation: Where it is in the public interest and does not represent a significant financial threat to the organisation, public bodies should make cartographic and georeferenced data freely available.
Postcodes
The introduction of postcodes will greatly aid the georeferencing of data. However, postcodes are designed to guide the delivery of mail. There is a unique opportunity in Ireland to enable postcodes to function as a set of mail instructions and as accurate georeferencing. In the other countries, such as the UK, this is not the case and there are a number of issues with using postcodes as a surrogate for georeferencing, such as having to use postcode centroids. It is vitally important that the best postcode system to meet mail and data requirements is introduced. The new Small Areas have been designed to nest inside every other administrative unit, have been tested for homogeneity, have established rules for splitting and splicing (to divide or merge areas depending on population falling or growing beyond certain thresholds), and have been adopted by OSI and CSO for data reporting. Given this rollout it would not make sense to devise a postcode system that is not harmonious with Small Areas. Recommendations: (1) postcodes should be either unique georeferenced address points or subdivisions of Small Areas; (2) postcodes are made available as a free-to-use look-up table (3) a representative of the Open Data group should be invited as part of the review process of Postcode tenders.
Institutional generation of spatial data and data governance
All government departments, local authorities, semi-state agencies and private industry that routinely generate data for use in their organisation needs to start georeferencing all their data as a matter of course at the scale of individual properties, either through Geodirectory or postcodes. This will require the collection of additional information fields at the point of generation and changes to databases to add address fields. Implementation of data capture, once tested, should be relatively straightforward. The requirement of adding fields – in terms of time or cost – should not be used as an argument to block georeferencing. This is a routine practice in the rest of Europe, will be required when postcodes are rolled out, and is needed in terms of becoming INSPIRE compliant. INSPIRE is the EU directive concerning spatial data infrastructures and public data and metadata standards across the European Union (see http://inspire.jrc.ec.europa.eu/). It stipulates that all EU countries must ensure that public data is georeferenced at a scale that enables localized analysis and comparison across the EU by 2019. Given the poor quality of addressing, there is every danger than Ireland will not be fully INSPIRE compliant by 2019 without tackling precisely the same issues as required for successful opendata initiatives. Even when georeferenced data is routinely generated there will be ongoing issues with respect to managing the generation and updating of data; processing the data in terms of cleaning, error checking and structuring; and outputting of data in useable formats. There are particular challenges if system relates to dynamic and streaming data from mobile public infrastructure (e.g., buses, trains) that aims to update in real-time, or if the system requires a response within a certain time period (as with fixyourstreet.ie). Organisations are therefore going to require sustainable systems of data governance. In a time of austerity, there may be resistance to implementing changes to data governance and institutional culture that open government entails. Recommendations: (1) all public bodies need to georeference their data as standard practice; (2) such data generation needs to INSPIRE compliant; (3) all bodies will need to implement strong systems of data governance; (4) consultation should begin with the Data Protection Commissioner on suitable thresholds for administrative data output.
Backcoding data
Given the poor state of addressing in Ireland, existing longitudinal datasets lack fine level georeferencing. It may well be the case that there would be much gained by retrospectively georeferencing these data so to enable fine-grained, spatially referenced longitudinal studies. Such retrofitting might be for a limited sample of the data at particular temporal points. At present, such longitudinal analysis is generally restricted to CSO data at ED and county level. It might be beneficial, for example, to examine the Live Register or planning permissions or house prices, and so on, at the housing estate or neighbourhood scale. Recommendation: (1) Every dataset should undergo a cost-benefit analysis as to the value of back georeferencing the data; (2) As part of the cost-benefit process Government Departments should undertake short pilot projects to explore the benefits from using spatially referencing datasets
Encouraging crowdsourcing
Crowdsourcing is the collective generation of media, ideas and data undertaken voluntarily by many people. The crowd works independently or collectively with respect to a particular issue. Fixyourstreet.ie is an example of crowdsourced data generation in Ireland where citizens can report issues in their area, such as road defects and vandalism, to local government, with the information being used to structure the work programmes of local authorities. To date, crowdsourcing with respect to georeferencing or mapping (such OpenStreetMap.org), or related to local government initiatives has been relatively muted. As such, Ireland is failing to exploit the power of crowds to address local and national issues. Recommendations: Examine the possibilities of harnessing the crowd to produce spatial data of use to citizens, state and business.
Threats to existing business
Creating widely available, free to use georeferenced and mapping data does provide some commercial threats to existing businesses. The business models of these companies are already under-threat from postcodes and the strategies being adopted by Google and opendata initiatives. They are going to have to adapt to opendata as it evolves and develop new ways to conduct business with open rather than closed data. Such a transformation has happened with respect to the music industry and is presently taking place in the book industry. Ultimately, Web 2.0 and evolving forms of digital exchange will affect all information and knowledge industries. Any threats to existing businesses should not be used as excuses not to take advantage of the significant benefits to Ireland of opendata through new economic opportunities, citizen empowerment, and more efficient and effective government. Recommendation: Do not stifle new opendata initiatives for short term protectionism that will ultimately fail and leave Ireland well behind the curve.
