C. GOVERNANCE MODEL, development strategy, risks & funding and sustainability - colouring-cities/manual GitHub Wiki

Under construction. Editing in progress by @polly64.

Introduction

This section describes the Colouring Cities Research Programme's (CCRP) governance, and explains why an academic governance model has been selected, and a collaborative maintenance model integrated. It also sets out governance challenges, potential risks and mitigation strategies and addresses funding issues.

Governance of infrastructure data platforms can be argued to be the most important dimension of data sharing platforms as this 'determines rules of operation of the platform, and facilitates and regulates the interaction of participants and the sharing of data and data services' Bühler et al. 2021. The World Economic Forum states that governance offers 'the possibility of collaboration directed by common principles as well as a space that limits human autonomy, a platform provides the opportunity to restrict unwanted behaviour and enable or incentivise desirable behaviour'. It can also reinforce environmental and social sustainability standards for the purpose of reaching common goals (ibid). Bühler et al. note the importance of protocols that are able to translate governance and design requirements into concrete procedures and source code; clarify and identify the specific roles and tasks of platform participants; explain relations and decisions associated with specific areas of governance; guide decisions towards shared values and objectives; operate as a common language for stakeholders; and aid, though non-technical descriptions, user/stakeholder engagement (ibid).

Summary of key Colouring Cities Research Programme (CCRP) governance challenges

Governance challenges for the CCRP, as identified during the development and testing of demonstration Colouring Cities platforms (2016-22) can be summarised as follows:

  • Developing a sustainable, low maintenance, trustworthy, governance model for CCRP platform rollout, suitable for countries operating in diverse economic and political contexts, able to create a step change in open access to high quality building attribute data at global level;
  • Co-developing and ensuring adherence to a common ethical framework and protocols across countries/academic institutions and ensuring that high ethical standards are maintained across platforms;
  • Developing mechanisms that allow awareness of risk of collecting specific data classes to be shared across countries, and addressed;
  • Sustaining a collaborative governance model as more academic institutions come on board;
  • Producing well documented open-source platform code able to be easily reproduced and co-worked on;
  • Producing an Open Manual, able to co-edited, that clearly sets our the CCRP's mission, aims and objectives, protocols, long-term development strategy, and methods;
  • Accessing and securing the highest quality comprehensive, open building footprint data, across countries, able to be updated on a regular basis as (these are the spatial filing cabinets/basic building blocks that allow data at building level, for stocks as a whole, to be located captured, collated, visualised and analysed, and from which other attribute may be inferred). Updates to polygons are also required to allow the scale and impact of change/energy and waste flows in the stock to be geolocated and monitored over time. Assessing security implications for footprint release, e.g whether building level footprints in CCRP platforms should be retained behind a fire wall, and implementing necessary action;
  • Maintaining quality across platforms
  • Maximising cooperation between international CCRP research and software engineering teams, and creating a non-competitive, inclusive, multidisciplinary, friendly, experimental and high quality research environment, that attracts and retains like minded research collaborators;
  • Demonstrating the way platforms can be easily independently managed by academic departments at country level, and developed as collaboratively maintained permanent national resources in conjuction with national academic networks that engage stakeholders from academia, government, industry, the third sector and citizens in uploading high quality up-to-date data;
  • Ensuring that the CCRP is clearly branded and differentiated from other platforms/ programmes using its open-source code, and that the security, privacy and well-being of platform users, building occupiers, and those involved in CCRP development, is always prioritised;
  • Minimising risks (see below) through system design and ongoing consultation, risk assessment, and monitoring.

Governance overview

Maximising CCRP system sustainability and resilience

Currently being edited The CCRP has been designed from the outset as a sustainability and resilience system that is highly adaptable, efficient and resilient. It does this:

  • By being a research-led programme, only open to international academic institutions specialising in building stock research to support UN SDGS. By their nature have common ethical and research standards and all work for the public good which acts as a first stage filter and reduces management- CCRP protocols.
  • By developing a programme that cannot ever be controlled by a single institution, or country owing to the fact that all code, data and platform development methods are released under open licences.
  • By developing an international academic programme where platforms are self managing and do require formal agreements with other partners(see also Zooniverse) and that allows academic partners to work at their own pace in line with their own research needs
  • By building on the success of distributed management systems for data collection projects i.e. Wikipedia and OpenStreetMap but increasing efficiency by combining crowdsourcing with automated collection, collation, generation and verification methods.
  • By facilitating the pooling of resources and expertise across coumintegrating develop the programme as an open knowledge system, thus s and supporting UN SDGs, existing framework in terms of protocols/research integrity, publications, research funding applications etc. which applies across countries.
  •  Initial set up only need footprints and CCRP open code and is inexpensive  to do simply requiring in-house engineering time from a RSE. This maximises the number of academic institutions able to join the network and means that RSE input is prioritized.

  •  The most important aspect of onboarding is the identification of platform value by the platform lead not only to their departmental/programme’s research and international networking  but also its longer term value as a national public open resource.  Capacity and interest to integrate platform development into an existing funded research programme  is essential. This helps filter out organisations that cannot see direct long term relevance to their research. It is essential that academic partners see the benefits to them of the CCRP engagement as being much greater than the value of the time they put in. This is a critical part of the process as the project is not sustainable if maintenance of these platforms stops just because one grant has been relied on and a grant has finished. Where maintenance of the resource is supported by a department, the academic institution, and a network of institutions across the country it is likely to survive. We focus first on value for academic needs in terms of data capture tools , as these are often relatively consistent across countries. Government policy will obviously vary and reliance on direct government funding is risky ( as opposed to an allocation for core database management across a number of relevant grants) as policy priorities may quickly change.

All countries will always find comprehensive collated built work in relation to energy reduction , conservation and planning stratgies, economic activity plans and disaster management etc . Many universities will also be looking for standardized data that can be analysed across countries- and big data able to aid AI and ML in gaining insights into patterns, cycles and likely trajectories in the context of sustainability and resilience. The platform allows for both standardized datasets and local datasets to be collected as long as they do not involve personal data collection.

All code we produce is open, meaning that anyone can set up a platform. However only academic institutions are permitted to join our network and can freely use the Colouring Cities name and logo and access our international meetings and workshops, joint papers and funding applications and international engineering group. Current partners see this as advantageous. If a partner wishes to drop out of the project they simply need to notify us and we will simply take them off webpages and off meeting lists. As all data must be released under an open licence another university in the same country may take over the GitHub repository and reuse collected data. The departing institution will continue to be credited on the Open Manual country specific page for its contribution. However this prevents wastage and one university within a country blocking progress for others.

We strongly encourage academic consortia to be developed within countries to create a distributed management network. In the Uk we are testing a university network across regions to support high quality data upload and moderation- (Australia has tested this at city level) . We are currently looking at two universities that would maintain the database together – which they could do as part of their existing funded programmes and through new applications.

The CCRP like Zooniverse is a global open science data collection initiative that tests a relatively new type of collaboration model within academia which draws from open data platform distributed management systems developed as part of projects such as Wikipedia and Open Street Map. The main differences is that we are a research led programme which is specifically designed to accelerate research to support Un Sustainability Goals and that focuseds on structuring and verification of specific types of comprehensive, microspatial daa and that uses diverse data capture methods to collect data rather than focusing on just crowdsouring . The system allows for new code can be rapidly deployed across platform, resources, engineering time and expertise can be quickly shared, and new countries can quickly come on board as part of existing research work. In Colombia it is undergraduate engineers who built the platform and who now have applied for their own grants and won these from the national mapping agency to continue their work.

We are consciously informal. We do not provide letters of support from Turing but letters of support can be organized from Hub members. The project is currently run from Turing but we are transitioning to greater engagement from partner organisations wishing to help coordinate Global Region Hubs. We have an academic steering group for Hub leads. The project structure is now resilient enough for any institution/country to drop out and it to still survive and to grow organically.

Platform design first begun to be tested as part of a multidisciplinary knowledge exchange project run in London in the 1990s which was co-built with citizens, with dedicated research beginning in 2014. The CCRP has developed incrementally, at low cost to maximise inclusivity and diversity of knowledge exchange and to make platforms sustainable. Everyone involved in the programme is a researcher so the pool of current expertise - with nearly 70 researchers, and shared expertised is great and will continue to increase. The project is still considered to be at an early stage of system design, with an increasing number of applications bring revealed as the system evolves. That it is being co-built and improved by the international consortium – both in platform core design and system oversight is critical to this process.

We aim over time to build a very large international research consortium of the highest quality committed to developing a global network of interoperable knowledge and data sharing platforms that provide free access to core built and natural environment data necessary to support UN SDGS.

1. CCRP Programme direction and international partner eligibility and self-filtering

**CCRP management and the Alan Turing Institute The Colouring Cities Research Programme was set up in 2020 at The Alan Turing Institute, the UK's national institute for data science and artificial intelligence. The Programme builds on research undertaken on the Colouring London prototype, at the Centre for Advanced Spatial Analysis, University College, between 2014 and 2019. CCRP research at the Turing has been funded by UK Research and Innovation, and by the Engineering and Physical Sciences Research Council's AI for Science and Government programme ASG. Though oversight is currently provided by the Alan Turing Institute, by 2025 distributed management of the CCRP across countries is is planned. The current role of the Alan Turing Institute is outlined below:

  • Knowledge and data sharing system design and project ethos
  • Management of access to core code repository on GitHub
  • Maintenance and development of CCRP webpages on The Alan Turing Institute and Github websites
  • Colouring Britain development and testing of new features and data capture methods
  • International partner engagement and co-ordination of knowledge/skills sharing
  • Provision of access for CCRP members to CCRP resources
  • CCRP protocol drafting
  • CCRP data ethics policy drafting
  • CCRP Steering group meeting set up and co-ordination, and Hub set up
  • Support and monitoring for CCRP academic publications
  • Accreditation methods

CCRP International Academic Steering Group and Global Region Hubs and CCRP International Research Software Engineering Group The CCRP Steering Group works informally to meet common goals set out within the UN's New Urban Agenda framework. The CCRP has begun its move towards to a distributed governance model. This involves voluntary contributions of time from international academic partners interest in co-building the global network and driving standardised building attribute data sharing across countries and regions.

Our aim as an informal academic consortium is to, collectively, produce a sustainable, supported network where institutions see significant benefits to collaboration; where costs of managing the network globally are shared and thus remain very low; and where there is sufficient flexibility to accommodate countries with diverse characteristics and resources to drive a step change in data, innovation and problem solving at national level. At the same time sufficient oversight and rigour to ensure CCRP global research aims, protocols and the CCRP's ethical framework are adhered to. In 2023 the Group was set up by the Turing to bring together CCRP academic partners interested in contributing voluntary time to leading on CCRP Global Region Hubs. Hubs have been set up to cope with CCRP expansion without grants being required; to address time zone issues impacting on the ability to attend partner meetings; and to enable academics, institutions and countries to work to together on issues common to specific geographic regions. No formal agreements are used. Platforms are funded through existing departmental/institutional research budgets, and supplemented by grants levered through platform delivery and CCRP networking. The CCRP European Hub was launched in January this year. All other CCRP Hubs, detailed above, are due to be launched by April. Additional Hubs will be set up as new geographic areas are represented.

  • Chair UK: Alan Turing Institute
  • Asia Pacific Hub lead - City Futures Research Centre, University of New South Wales, Australia: Professor Chris Pettit and Dr Matthew Ng
  • N. America Hub lead- Canada Excellence Research Chair (CERC) in Smart, Sustainable and Resilient Cities and Communities, Concordia University, Montreal: Professor Ursula Eicker and Dr Oriol Gavalda
  • European Hub lead- Leibniz Institute of Ecological Urban and Regional Development Research Data Centre: Professor Martin Behnisch, Dr Robert Hecht and Dr Hendrik Herold
  • Middle East Hub lead - University of Bahrain, Urban Housing Lab. Dr Wafa Abdulrahman Ahmed Alghatam
  • Latin American Hub lead - Dr Fernando Benitez, St Andrew's University, UK

CCRP International Academic Partners and Protocols CCRP academic partners voluntarily sign up to common protocols and aims agreed by the Steering Group. Each collaborating institution has specific expertise in the built environment and views Colouring Cities as adding value to its long-term research strategy and offering new opportunities for international collaboration, multidisciplinary working, and new funding streams, with CCRP databases relevant to many areas of academic research including planning, housing, transport, economic regeneration, retrofit, and climate risk to understanding the spatiotemporal distribution of deprivation and ill health. All partners are already be involved in research undertaken to help increase the sustainability, quality, efficiency and/or resilience of building stocks, and each sees the value to providing standardised open data across countries to support urban problem solving and net zero targets at the global scale.

Each participating institution takes full responsibility for all aspects of its own use of CCRP's open code, and for the legal, ethical, financial and management oversight of CCRP platforms at country level. Each academic institution involved will also, by its nature, be subject to a separate set of rigorous standards regarding academic integrity and ethical behaviour.

Application to become a CCRP platform partner, and to gain access CCRP resources other than open-source code, currently involves initial informal discussion with the Alan Turing Institute regarding the challenges, risks and opportunities that collaboration may bring. Academic partners are strongly encouraged to identify clear benefits of long-term involvement for their own research, and their departments and institutions before joining. All CCRP partners are experts in specific aspects of the building stock and voluntarily share skills and knowledge to enrich and sustain the programme as a whole, help meet UN SDGs and maixmise benefit for individual countries and institutions involved. Interested institutions/consortia are simply asked to confirm in writing a commitment to CCRP protocols and research goals and to provide evidence of research institution status and multidisciplinary expertise (as shown in Section M). Independent collaboration between CCRP partners on specific areas of development is strongly encouraged.

The time frame for platform development is open, however a clear demonstration of help-in-kind support/funding for live demonstration models, and a commitment to working to set these up within a 6 months period is expected. CCRP partners receive access to free CCRP resources and meetings coordinated by Turing, and are immediately included on the Turing website website alongside other international partners and on the Manual's international partner list.

The development strategy does not involve direct approaches to institutions, unless research programmes of particular relevance are flagged, as the CCRP programme is specifically designed to develop organically and to be self filtering. Academic departments interested in joining the CCRP, and signing up to CCRP protocols, must have the capacity and energy to identify funding and software engineering support to set-up demonstration platforms, and begin stakeholder engagement themselves, or else it is unlikely that platforms will survive beyond the extent of the initial grant.. Institutions willing to make these efforts, without prior funding, are considered likely to have the drive, imagination, skills and commitment necessary to become effective Colouring Cities platform leads, and to have the capacity to sustain platforms and help drive collaborative research programmes as permnantly evolving public resources.

All CCRP international academic partners are registered academic institutions. Only one academic institution or research consortium per country is able to be represented on the CCRP. CCRP academic partners are wholly responsible for all aspect of individual open data platforms including set up, technical maintenance, management, funding, legal and ethical oversight and sustainable development, and for the visualisation and open release of building attribute data for use by stakeholders and the general public. CCRP partners commit within CCRP protocols to maximising trust in platforms, optimising quality and coverage of data, supporting stakeholder engagement, and promoting high standards of data ethics. Partners also commit to driving and supporting relevant research collaborations at national and local level, necessary to support a sustainable funding model (see 'Costs and Funding' below).

CCRP academic partners are required to hold research expertise in at least one of the following areas to be eligible as platform hosts: planning, building construction, building design, energy and retrofit, building conservation, building quality and performance, typology classification, building history/lifespans and demolition patterns and/or GeoAI relating to building attribute data. Involvement of internal computer science departments/research software engineers and students forms an essential part of CCRP strategy. CCRP academic partners must also be interested in drawing on multidisciplinary knowledge of the stock from diverse stakeholders, and have extensive of experience of the analysis and/or modelling of building attribute data to advance CCRP research goals.

International academic partners operate as national platform hosts,and together form the CCRP international academic network. Each sign up to CCRP protocols and the CCRP's Ethical framework. CCRP partners with platforms operational for over a year, that have also given voluntary time to support set up and management of CCRP global region hubs are also encouraged to co-work with the Turing on long-term development strategy. CCRP platforms also offer opportunities for individual research institutions to bring staff and students together from diverse disciplines to work on a public, visual database combining knowledge from science and technology, the arts and humanities, and from communities to support United Nation's Sustainable Development Goals.

The initial task for CCRP academic partners is to set-up, and manage Colouring Cities demonstration platforms - at town, city or country level. The first step in this process involves securing and visualising building footprints, ideally from national mapping agencies, which form the basic building blocks. (These do not have to be open but must be able. to be visualised and used to capture and colour data); the second, combining CCRP code with these footprints and creating the Colouring Cities platform and database; the third, managing platforms in line with the CCRP protocols and ethical framework; the fourth, engaging stakeholders from academia, government, industry, the third sector, and citizens, in the co-creation of the database; the fifth expansion of the platform to national level and showcasing of data applications; the sixth, co-working with other in the CCRP international network to enrich CCRP code and to drive multlidisciplinary international research collaborations.

Academic partners may relinquish CCRP membership (and associated Colouring Cities branding and resources and network access) at any time and still continue to use Colouring Cities open-source code to run their open databases, even though they will no longer have access to CCRP branding, domain names or other resources. It is the task of the CCRP as a whole, and essential if it is to survive and flourish, to demonstrate the value of ongoing membership of the branded academic research initiative, rather than to use CCRP code outside this, and to help potential partners identify both challenges and opportunities.

The role of CCRP international platform hosts is summarised below:

  • Securing comprehensive open building footprints to provide the means by which data on stocks may be collected, collated and visualised and ensuring footprints are managed in a responsible way.
  • Setting up demo models using CCRP open-source code at country level with a view to creating national databases
  • Adhering to CCRP protocols (including in relation to prototype testing, research expertise, ethical use of data, accessibility and inclusivity of platforms and databases, and the development of constructive, productive, generous spirited collaborations to advance research goals) and maintaining high ethical standards.
  • Funding and managing national open data platforms
  • Assessing and addressing liabilities resulting from platforms
  • Developing platforms as agile, low maintenance open national resources designed to operate as permanent resources adapted over time
  • Engaging national stakeholders and securing logos (i.e. trust marks) from relevant national public bodies
  • Advancing research questions through the capture and analysis of data
  • Sharing knowledge across countries on technical and research issues through the CCRP academic group
  • Co-working on papers and joint funding bids with CCRP
  • Co-ordinating upload of large-scale open datasets contributed via CCRP national platform data portals
  • Ensuring collaborators are credited and thanked at national level.

3. National/Regional CCRP academic networks

To maximise the quality and quantity of data available within CCRP national platforms, regional academic input is also required. A model for national academic networks is therefore also being tested, encouraging a university or group universities within a region/city to drive data upload, and moderate data, as part of their research. Moderation of all bulk uploads is required. The idea is to build in academic oversight, and involvement in data capture and verification both at the macro and microscale; to encourage cross sector partners already working with academia (in government, industry and the community) to collaborate on national databases, and to test and showcase applications of CCRP data in diverse areas of research. Regional/city academic networks are currently being tested in Australia (at city scale) and Britain (at regional scale). Involvement is designed to be funded initially through existing research programmes managed by lead institutions, later supplemented by new grants available from across sectors that encourage greater access to building attribute data.

The model allows for all universities within a country to benefit from access to a unique, free, ready-to-use open national platform, to which they can contribute to and access over 100 data building attributes from. It also provides access to a unique system that tests four methods of data capture and feedback loops between these (i.e. Crowdsourcing of data from diverse sectors/communities at building level; streaming of government data; collation, cleaning and upload of existing open bulk datasets); is relevant to many areas of academic research from planning, housing, transport, health, economic regeneration, retrofit, and risk assessment relating to climate change to the spatial location of deprivation and ill health; offers opportunities to bring staff and students together, from diverse disciplines, to work on a public, visual database collating knowledge from science and technology, the arts and humanities, and communities to meet UN Sustainable Development Goals; and by opening up new research opportunities. For academic leads overseeing data input within regions and cities, grant new opportunities are likely to exist.

4. Public institutions, government and third sector contributors

Collaboration between public institutions/bodies - whether in research, government or the third sector - and CCRP partners forms a critical component in the development of collaboratively maintained national databases, where responsibilities and costs for database management and quality are shared. Integration of national public datasets is most likely to be managed by international CCRP partners responsible for the technical and strategic development of platforms at national level. Regional academic networks are most likely deal with public data upload at regional level, though may help with national upload for specific datasets lying within their area of expertise. More formal collaborations between CCRP hosts and stakeholders may arise through this process owing to extensive opportunities for the public institutions to improve, integrate and publicly share their datasets within an integrated free public resource, at very little cost.

Advantages for the CCRP regarding collaboration with public sector bodies includes: increasing access to large-scale high quality datasets relevant to specific CCRP categories; speed of upload of data as a result of collating datasets; broadening and reducing costs of improving and update specific types of dataset through co-working using crowdsourcing features; opening up opportunities public institutions to apply for grants to connect their knowledge and resources to an central open resource; freeing up CCRP partner time to focus on platform infrastructure and technical management, on improving open code with CCRP partners, on testing diverse data capture methods and feedback loops, and on working on related collaborative CCRP analysis and research initiatives. The distribution also maximises opportunities to improve database efficiency and quality and reduce platform management costs, as well as improving coverage, quality and speed of upload.

Public institutions are encouraged, where of interest/benefit to them, and where funding exists or is able to be applied for by them, to:

  • Contribute large-scale, cleaned open datasets via CCRP moderated portals
  • Encourage staff to directly enrich and verify relevant datasets
  • Monitor and moderate specific datasets where possible (e.g. historic buildings data, housing data, energy data
  • Recommend new datasets for inclusion and data formats/standards for specific data types
  • Demonstrate data applications using the planned CCRP Showcase section (see Section I).

5. Historic environment sector

Input of historic environment sector specialists is of primary importance in the development of CCRP databases. These specialists include building historians, conservationists, urban morphologists, members of national amenity societies, and of local preservation and civic societies. The sector's expertise is unique in that it covers all three areas of interest addressed by CCRP platform i.e. relating to stock composition, building performance and short and long-term stock dynamics. Long-term knock-on effects of change, avoidance of finite resource depletion through repair and reuse, tracking of survival, and the promotion of incremental adaptation to maintain and create diverse, sustainable and resilient urban systems, are all well understood. Many individuals within the sector are also commonly involved with voluntary activities relating to conservation.

Historic environment specialists are encouraged to:

  • Contribute data on building age/construction date, demolition and lifespan data, typology characteristics, original construction team, planning controls and designation, materials and construction systems and typology performance
  • Verify manually contributed datasets
  • Verify datasets generated computationally, especially those relating to building age and typology
  • Co-work with the CCRP academic network on rules for typology geolocation
  • Co-work s a sector live on CCRP maps to fill missing data in.

6. Commercial sector

A significant proportion of building attribute data is held within countries by the commercial sector. In particularly in areas such as building insurance, property development, banking/mortgages, property search/information services etc. Increasing involvement by commercial companies such as Microsoft and Google in the generation of open building footprints at global level also indicates the increasing interest in commercial opportunities for data capture.

The CCRP provides the commercial sector with permission for third party use for open data to help drive innovation in analysis, models, products etc. to support the sustainable improvement of stocks and net-zero goals. Competition is encouraged through commercial leaderboards, allowing companies to demonstrate the scale of their contribution of data for the public good, within or outside corporate responsibility programmes. The more data companies can contribute data to CCRP platforms, the bigger the pool of free high quality data for the commercial sector (and others) will be.

Commercial logos are only recommended to be included on CCRP platforms on the dedicated commercial sector leaderboard page.

7. Citizens and Schools

Citizens and schools are also encouraged to enrich and verify databases through building level voluntary contributions. Of particular interest is input on building typology performance from the citizen's point of view. Though it is not anticipated that significant amounts of data will be collected from citzen's (other than those involved in historic buildings and building conservation) other than for their homes, opportunities exist for schools to make a considerable impact in specific categories such as land use at local, particularly where data collection can be coordinated as part of a national of regional initiative. Data for the first land use survey in Britain, undertaken in the 1930s by the London School of Economics, was primarily crowdsourced from schools.

Citizens and Schools and Community Groups are encouraged to:

  • Add to and enrich datasets at building level on their homes and local areas
  • Verify local data
  • Assist in colouring in gaps on maps
  • Become involved in contributing data able to support net zero goals

Main governance risks, and risk mitigation strategies

Potential governance risks, identified to date, and risk mitigation strategies are set out below. Additional risks identified will be added to this list as and when identified by CCRP partners. Responsibility for risk assessment and for mitigation and monitoring lies with individual Colouring Cities partners.

  • Risk 1: Inadequate governance, management skills, research expertise or infrastructure to set-up and sustain high quality CCRP platforms.

  • Risk 1 Mitigation: Selection of trusted public research institutions only as national CCRP platform hosts; rigorous requirements for CCRP membership: e.g. must have relevant research track record, multidisciplinary expertise, ability to fund and set up demo platforms; detailed advance discussion with potential partners on challenges and risks, and assessment of capabilities/interest; clear protocols and ethical guidelines for hosts; clear messaging to prospective hosts that platforms are designed to operate as permanent public databases and research tools able to built on and adapted in future and are not one off projects; regular CCRP platform host meetings allowing potential risks and mitigation strategies to be discussed..

  • Risk 2: Difficulty in accessing reliable, high quality software engineering expertise (essential for set-up and technical maintenance and enrichment), and high costs of engineering time.

  • Risk 2 Mitigation: Requirement by CCRP of prioritisation by academic partners of appointment of software engineering teams; early exploratory work by CCRP partners regarding possible collaboration with national centres of excellence in software engineering; regular meetings of the CCRP international engineering group to facilitate skills and knowledge sharing, support collective problem solving and prototype improvement, and help reduce engineering and platform maintenance costs across countries.

  • Risk 3: Failure of CCRP partners to access comprehensive, high quality footprint data to enable spatial data to be captured, collated and visualised at national scale, or to see updating as a priority for maintaining database quality.

  • Risk 3 Mitigation: Requirement by CCRP of prioritisation by academic partners of open footprint negotiation/agreement with national mapping agencies, to access comprehensive mapping data of the highest standard (even if initial permission is not for full release as in UK model); knowledge sharing between international CCRP partners on alternative options if negotiation fails e.g. temporary use of OpenStreetMap polygons; use of CCRP examples to demonstrate value of national mapping data to governments to help drive their release.

  • Risk 4: Failure to create successful collaborative maintenance systems at national level and to facilitate cross sector/disciplinary working with stakeholders, essential to develop and maintain high quality data platforms at low cost, and to optimise data quality and platform resilience.

  • Mitigation: Use of an academic governance model that allows academic hosts to provide a free, trustworthy, canvas on which diverse stakeholders can work - whether these be public institutions, commercial companies or citizens/citizen groups - at a time, and in a way, that benefits them. Clear accreditation of all contributors on the platform, use of public body logos as trust marks and inclusive interface design viewed as critical to attract the widest range of stakeholders possible.

  • Risk 5: Potential collapse of databases owing to early rapid expansion, and reliance on maintenance of costly, unnecessarily large, centralised development teams;

  • Risk 5 Mitigation: Strong recommendation by Turing to platform hosts to develop incrementally, in collaboration with stakeholders, and to manage programmes with small core teams, but many institution partner collaborations.

  • Risk 6: Loss of trust in the CCRP owing to poor platform hosting, or lack of progress on platform set-up by CCRP partners

  • Risk 6 Mitigation: As Mitigation for Risk 1 plus inclusion in CCRP protocols that Turing may remove CCRP platforms from its CCRP website where academic hosts are not engaging with the CCRP group/advancing platforms in line with protocols, and/or have not communicated with Turing for over six months.

  • Risk 7: Loss of trust in the CCRP owing to applications of code and data with negative impacts outside the CCRP programme

  • Risk 7 Mitigation: Strong, clear and tightly controlled branding of the research programme and Colouring Cities interface design and logo; clear differentiation of the project from other initiatives; demonstration of accessibility, transparency, inclusivity, openness and ethical behaviour by platforms, and association of the Colouring Cities brand with these; clear messaging of the aims, potential and limitations of the platform, so that the public and stakeholders/platform users understand exactly what it does, how it is doing it, what they are contributing to, issues relating to it, and how they can support its public research goals.

  • Risk 8: Change in priorities of CCRP platform hosts over time leading to platform collapse and knowledge/database loss;

  • Risk 8 Mitigation: Acceptance from the outset that some platforms will fail or, in time, consider relocation and integration in say government run platforms; clear identification of benefits of ongoing CCRP membership; live information on platform successes - initially through individual Wiki pages (Section M) and later through a new CCRP webpage operating as a portal to all platforms; provision of unique resources to CCRP platform hosts to encourage ongoing interest and engagement including opportunities to analyse data with AI and machine learning specialists; self-filtering mechanisms included in CCRP partner application process to help increase probability of long-term engagement/commitment.

  • Risk 9: Failure of CCRP hosts to secure sufficient funding

  • Risk 9 Mitigation: Includes tailoring content to maximise funding steams, developing new research networks, and maintaining low maintenance platforms and core project teams. See also 'Funding' below.

  • Risk 10: Failure to supplement crowdsourcing with automated data capture methods required to produce large scale datasets of a scale and quality sufficient to attract and engage expert stakeholders in maintaining and enriching relevant datasets

  • Risk 10 Mitigation: Inclusion of this as a research goal signed up to by CCRP international research partners; sharing and testing of techniques across the CCRP group.

  • Risk 11: Failure to protect the security of platform users or of occupiers of buildings to which data refers

  • Risk 11 Mitigation: Inclusion in protocols that it is the responsibility of academic partners/platform hosts to prioritise data ethics and the security and privacy of platform users, building owner/occupiers and CCRP staff; knowledge sharing on potential security issues and liabilities within the CCRP international group; assessment of datasets against agreed data principles (see ethics); capture of free text only where essential; transparent development through GitHub; rigorous filtering and monitoring of feedback fora by platform hosts to minimise malicious comments.

  • Risk 12: Failure to provide adequate information on platform interfaces to protect platform hosts or contributors from liability in terms of issues with data inaccuracy, openness and use.

  • Risk 12 Mitigation:* Protocol inclusion that it is the responsibility of academic partners/platform hosts to check potential liabilities relating to platforms with their legal departments; provision of clear disclaimers on platform interfaces stating responsibility for assessment of data accuracy lies with users; checking of any liabilities by individual countries with host academic institution's legal teams.

Choice of academic governance model with collaborative maintenance over other governance types

The CCRP governance model has been arrived at following six years of experimentation with the Colouring London prototype platform, first at the Centre for Advanced Spatial Analysis, University College, London, and, since 2020, at The Alan Turing Institute. From 2020 the model has also been influenced by findings from the set-up of international platforms.

Summary of advantages of academic governance

In the context of governance challenges, academic institutions are the only bodies with the infrastructure, principles and drivers necessary to provide suitable oversight and direction for the CCRP platforms, and with the regulatory oversight, and longevity, to be viewed by the majority of stakeholders as reliable, trustworthy and effective hosts for permanent CCRP platforms. Universities and public research institutes are subject to high levels of public scrutiny and require researchers to follow ethical guidelines with regard to how their research is carried out, how information is collected and used, and how findings are published and credited. Academic institutions have been identified as ideal hosts for open building attribute data platforms owing to their exceptionally high trusted status compared to other types of organisations (Universities UK, 2021). Universities and other publicly funded research institutions, such as Turing are not-for-profit, politically neutral bodies which are also repositories of expert knowledge; are run to support the advancement of knowledge for the public good; have rigorous built-in procedures relating to data ethics and data use; are commonly required to to find ways of engaging the public in research (ibid.) and have long time-horizons. These are all necessary to help facilitate partnerships required to create long-term, complex collaborative maintenance systems in which different diverse sectors and disciplines contribute, update and monitor building attribute data within their areas of expertise. This provides a common language and code of conduct for researchers that already bridges national boundaries. The type and range of expertise required to support the CCRP's research goals is also known to exist within academic research programmes. Universities look to work independently of political or commercial interests; understand the importance of testing models and of evidence-based decision making; are experimental and push research boundaries, and have the capacity and interest to test diverse data capture approaches to effect a step change in the scale, range and quality of open data on buildings, and national stocks. All will be experienced in fundraising for research, working across institutions and countries, and working with diverse audiences/partners. Public research institutions also possess the necessary infrastructure, and driver, to both set up and maintain open data platforms as permanent research and teaching resources and tools, for the public good, and understand how hosting of international research initiatives potentially offer wider benefits in terms of student interest, funding and PR. Initial testing, at UCL and the Alan Turing Institute, has indicated that national research institutions or academic consortia may be better placed to act as Colouring Cities hosts than individual universities owing to the ability of research institutions to create less competitive environments.

Issues with alternative governance models

Government, commercial and third sector governance models were also assessed but found to be significantly less well placed to address governance challenges and risks. The fact that governments do not already manage integrated databases on building level data on building stocks, despite the relevance of building attribute data for government, industry, academia and the public sector, and to meeting global sustainability goals illustrates the scale of issues existing in joining up thinking within departments with large, focused delivery briefs. In the UK for example sharing of property tax data, the richest source of data on stock composition for many countries, between the tax and housing departments continues to be problematic. Platforms must also be as free from political influence as possible, and from being required to meet short-term political goals. Though in the commercial sector, necessary technical and research capabilities and funding to experiment and innovate in specific areas of platform design clearly exists, the sector is not suitable to oversee public resources of this kind. CCRP platforms are non-profit research initiatives designed to be supported through collaborative maintenance systems to meet research goals specifically designed to benefit society as a whole. However collaborations with the commercial sector are important to drive innovation - with the popularity and value geospatial data greatly advanced by it, though decisions to access to data, made freely available on commercial platforms does need to be considered with caution, as such access may be removed or altered at any time. Lack of public trust in tech companies also may inhibit input from communities with increased commercial engagement in open crowdsouring projects already raising concerns.

In terms of the third sector, the remit of individual public institutions involved in the built environment i.e. in housing, architecture, surveying, engineering, planning, heritage etc.) is too limited in scope to manage the range of data types (identified as required by stakeholders to improve sustainability, resilience, efficiency and quality), with thee bodies also not set up to manage the diversity of research collaborations necessary to address key research questions.

Open Foundation models

The 'Open Foundation' governance model, as used by Wikipedia and Open StreetMap (OSM), has proven extremely successful in managing the growth of global open data/knowledge exchange platforms and databases that are sustained and underpinned by complex collaborative maintenance systems. Both the aforementioned examples have demonstrated, over now many years, that where open data of popular interest are able to be contributed to and edited on platforms by the crowd, and where indicators of reliability (such as sources and history of edits) are also included, incentives to continue to improve, maintain and expand databases will grow, in turn increasing the usefulness, reliability and trustworthiness of platforms and public interest in them. Both Wikipedia and OSM have managed, using this model, to effect a step change in access to specific kinds of open data. OSM (https://wiki.openstreetmap.org/wiki/About_OpenStreetMap) is an open mapping initiative, set up in 2006, that initially focused on street network data but which has more recently made significant inroads into the collation and capture of open building footprint and land use data. It is run as a non-profit Foundation and is run by a board of directors who are voted in by OSM members, and an advisory board. OSM currently has over 8 million users. The OSM Foundation's main role is to oversee the OSM server and interface, accept and allocate donations, and engage with working groups (https://wiki.osmfoundation.org/wiki/About). Wikipedia is an open online encyclopedia that currently has over 43 million users and 1.8 billion visitors monthly. It is managed by the Wikimedia Foundation which was was established as a nonprofit in 2003 and hosts 13 collaborative knowledge projects. Like the OSM Foundation, the Foundation's role is to maintain servers, build relevant software, and collect and distributes donations to collaborating groups. It also actively seeks partnerships and has over 500 staff and contractors, a board of trustees and leadership team (https://wikimediafoundation.org/our-work/).

The CCRP, like Wikipedia and OSM operates, is a knowledge sharing, collaboratively maintained initiative that also operates at international level and has strong community roots -originating in the 1990s as a free physical knowledge crowdsourcing initiative about the building stock that also experimented with the public visualisation of geospatial data. In 2019 the Colouring London prototype was named by the Open Data Institute (ODI) as an example of a successful collaborative maintenance model, alongside Wikipedia and OSM, owing to collaborative maintenance being a key component of its design. The ODI defines common characteristics of these platforms as:

  • Establishing and nurturing a community that collaborates continuously to solve a problem.
  • Producing a meaningful and quality database.
  • Making data more accessible in a way that can be useful to as many people as possible (ibid.).

However unlike Wikipedia and OSM the CCRP has been set up to answer a number of specific research questions including those relating to the impact of of open databases on stocks on sustainability research and practice; how feedback loops between automated and crowdsourced methods of data capture could best be created to improve data accuracy, and to the potential effectiveness of platforms as stock auditing and performance monitoring tools. The collaborative maintenance system being developed is more structured with networks of universities and other public institutions being brought together as primary editors and moderators, and also involves collaborative working on the academic programme and code itself.

Much has been, and continues to be learned, from open database foundations and initiatives in relation to:

  • methods of successfully operating collaborative maintenance systems at global scale;
  • speed with which loosely managed, high value open data platforms can be built, and managed, at low cost;
  • scale of demand for open data platforms
  • in terms of Wikipedia, use of endowment funds to sustain open databases in perpetuity;
  • importance of specific types of open licences and of open data protocols;
  • value of specific interface features in supporting stakeholder (such as the OSM edit history feature);
  • the importance of non-technical access for users to increase users numbers/expert input;
  • lack of diversity of editors for Wikipedia and OSM;
  • the value of platforms as live spatial data capture tools, as shown by OSM, able to aid disaster management.

Citizen Science governance models

The closest governance model identified, in relation to CCRP operation, is that used by the citizen science project Zooniverse. Citizen Science involves participation of non-scientists in scientific research with the online crowdsourcing knowledge, skills and data from non-scientists being a popular method of engagement. Zooniverse is a collaboration between the University of Oxford, Chicago’s Adler Planetarium, and the University of Minnesota – Twin Cities (UMN) in which crowdsourcing is used to support a range of areas of scientific research. Zooniverse currently involves over 2 million participants and hundreds of researchers from diverse research teams, and, like the CCRP, brings together expertise in academic and scientific research, public engagement, and web platform design. Similarities also exist in academic oversight, informal involvement of research institutions across countries, harnessing of institutional support, infrastructure and public status, coordination of grant funding, exploitation of user skills and application of data to advance scientific goals. The key difference in the CCRP and Zooniverse models lies in the CCRP's interest in supporting and initiating many types of research relating to the socio-cultural, economic, environmental well-being of cities and urban areas, to meet UN global sustainability goals, by collecting many types of attribute data relating to a single, specific type of object; i.e. buildings). The CCRP also requires a governance structure able to support testing of open-source code, and collaborative analysis of captured data by international research collaborators, and to oversee protocols for the rollout and management of permanent international open database/platforms, with high ethical standards, able to be used in scientific, and other types of, research.

Funding streams and resilience

Developing a low maintenance, sustainable governance and management model for the CCRP and associated platforms is seen as essential to allow the CCRP to grow and flourish, and in doing so to increase open knowledge, and open data sharing and availability, at global scale.

Key methods tested to minimise costs, and maximise the lifespan, effectiveness and resilience of platforms include:

  • Keeping core development management teams small, but research collaboration networks big.
  • Aiming to attract help-in-kind contributions of time and data at national level of of value at least triple that of core platform running costs.
  • Setting up and managing low maintenance demonstration platforms within academic departments already involved in building stock research, where platforms are immediately relevant, enabling some overheads to be absorbed;
  • Using demonstration platforms to show the benefits of CCRP platforms to research funders;
  • Sharing software engineering time across countries;
  • Sharing strategies across CRRP teams for increasing platform efficiency;
  • Distributing fundraising responsibilities by building strong partnerships with bodies interested in leading funding bids for research where data on the composition, dynamics and/or the performance/quality of the stock are needed - where CCRP hosts can operate as grant partners and co-investigators;
  • Consulting and engaging with multisector stakeholders to increase help-in kind support and highlight potential funding opportunities;
  • Diversifying in terms of funding sources; using multi-source funding to increase project resilience;
  • Integrating platform as much as possible into teaching and research programmes, and maximising voluntary opportunities students to engage;
  • Being transparent regarding funding sources and publishing funding information on country pages (see Section M) to help other CCRP partners explore similar types of sources for platform set up and management;
  • Creating opportunities for joint international grant applications with other CCRP members where demonstration platforms have been set up;
  • Including CCRP Protocols requiring prioritisation of sustainable funding models;

As the stock represents such an important part of national infrastructure, is a major emitter of greenhouse gases and is a city/country's most important socio-economic and cultural resource (Bradley and Kohler, 2007), significant opportunities exist for CCRP partners to tap into a diverse array of funding streams. The time consuming nature of PI time securing funds for the Colouring London prototype in a highly competitive grant arena, as well as the need to sustain platforms at as low a cost as possible. Platforms are well placed to exploit diverse smaller scale funding streams in collaboration with research partners and where management teams are and overheads are kept small/low. Grant awards from specific sectors are also important in increasing trust and interest from specific stakeholder groups, demonstrating the value of specific content to these sectors and to increasing likelihood of their long-term, in-depth engagement.

Fundraising for Colouring London/Colouring Britain and CCRP management has become considerably easier as the platform and its networks have become more developed. Over £900,000 has been secured to date, mainly for the Colouring London prototype. In the UK the current strategy is to keep the Turing team running the CCRP and prototype development, small and agile (at around 2 FTE posts), and maximise to additional research input through in-kind help from collaborations, and joint grant applications led by research partners. Turing's core team is currently funded by a c£200,00 UK AI and Science for Government grant which covers direction of the international and national programme, and development and testing of Colouring London and Colouring Britain open-source code, and prototype management(see also Colouring Britain for grants received 2016-2020). Information on funding for individual country platforms can be found in section M.

General sources of funding relevant to CCRP platform maintenance and development:

  • Academic grants (existing and new) in areas including: open data, urban analytics, urban science and the mathematics of cities, data science, data visualisation, smart cities, energy, housing, planning conservation and heritage, sustainable cities, risk assessment, disaster management, AI and machine learning, digital twins
  • Research innovation grants for example relating to the development of open data systems and open-source code; opening opportunities for the use AI in urban problem solving (socio-economic/environmental), and experimenting with new methods of public engagement in sustainable city development
  • Government grants relating for example to housing, planning, infrastructure, energy, sustainable cities, heritage, urban vitality, health, education, community well being, community engagement, data access, innovation in technology;
  • Non-profit grant giving institutions and charities involved, for example, in housing, energy, building design, construction, conservation and repair, planning, education, innovation in technology, sustainability, and community engagement in planning;
  • Grants from educational trusts set up by the construction/property industry.

funding

Additional note: Wikipedia is particularly interesting in terms of the scale of funding secured for its collaboratively maintained open database. Wikimedia has an endowment fund of c$100 million which is used to provide a base level of support for Wikimedia projects 'in perpetuity'. In addition to this Wikimedia raised $162 million in 2020/21 and holds total assets for that year of $231 million, which exclude endowment figures, compared to $80,000 for its setup year 2003/4. Its annual expenditure for 2020/21 was £111 million compared to $23,000 in 2003 (https://en.wikipedia.org/wiki/Wikipedia: Fundraising_statistics).