Data ethics are described by the Open Data Institute (ODI) as a 
" A branch of ethics that evaluates data practices with the potential to adversely impact on people and society-in data collection, sharing and use." 

Ethical use of data brings about trust and helps allow data to work for everyone.
The Colouring Cities Research Programme uses the ODI Data Ethics Canvas to help identify and manage ethical issues throughout the lifecycle of its prototype platform Colouring London. 
 
As part of the process of development, existing, and new features within the platform are checked against the questions posed by the Ethics Canvas. First stage responses to core questions are given below. 

Colouring London collects over fifty subcategories of data to support research into the sustainability of London's building stock. These relate to building location, use, type, age and history, size, materials and construction, sustainability (including energy rating and estimated lifespan), design/construction team, planning/ designation/demolition status, streetscape/green context, whether the building is community owned, and whether the user thinks it contributes to the city.
 
Our data are available for download as open data from our platform. Most data we are collating, collecting and/or generating relate to physical characteristics of building, already able to be seen from street or satellite images. Much of this information is also already held within government or commercial databases, though in many cases these are restricted to the public and academia. Some government datasets, such as building designation and protection and energy rating are already publicly available.  Detailed information on London's buildings is also provided by the commercial sector with images of the interiors of homes, their price and sale history now commonly available on property websites.  

Colouring London doed not collect personal data, other than optional emails needed to enable users to reset site passwords. We actively discourage users, on our 'sign up' page, from contributing even their real names.  Though it is helpful to understand what sectors and disciplines and groups our users are coming from to help us reach and ensure relevance to as wide an audience as possible, we believe requests for information for this purpose should, if introduced, be optional only, with minimal information asked for. We will continue to explore how this issue can best be addressed by working and consulting across sectors, disciplines and community groups.  
 
Our sign-up agreement also tries to be as transparent as possible. It emphasises that when users make a contribution to Colouring London they are creating a permanent, public record of all data they add, remove, or change; that the database will record the username and ID of the user/editor, along with the time and date of the change and that all of this information will be made public through the website and through bulk downloads of the edit history.
 
 Data are gathered in four main ways.  Firstly by identifying and collating existing datasets held by central and regional government bodies and other organisations. Secondly by harnessing knowledge held within the community through crowdsourcing at building level, whether this be from, for example, building professionals, local councils, local amenity societies, building users or schools. Thirdly, through large-scale computational data generation programmes, with results encouraged to be verified by local experts as well. Live streaming of planning data will also be tested.
 
Our job is to bring together and visualise data on the building stock that is currently high fragmented, restricted or unavailable, to make this more accessible and increase data accuracy through inclusion of sources and verification.  Some data may derive from observation of the building itself, some need to be extracted from historical texts and some come in the form of ready-to-go datasets comprising city wide spatial statistics.  Users are informed on sign up that data cannot be accepted on the site where any restrictions to its open release may apply.Our 'Community' section differs slightly in that it also asks users' how well buildings work and whether they contribute well to the city and/or local area.  

The platform is also being designed as a collaborative data maintenance project as described in the ODI's handbook at https://collaborative-data.theodi.org/, with specific datasets encouraged to be added to, verified and updated by specialist sources (e.g. historians for age data, government energy specialists for energy data, and communities for workability data). Our stewarding structure is in the process of being developed.

Data on buildings are collected at building and property level (a property representing  either the whole, or part of, a building). Significant questions currently exist in the UK with regard to access to address/location data and substantial work is still required on how data can best be matched to footprint polygons.

To help address issues of accuracy and bias a number of features are being included. Each subcategory has a source box, a verification button, with a  query button planned to enable problems that cannot be addressed within the editing system to be raised. Moderated dropdown options plus links to allow references to sources and routes to further information are also included. Easy to access edit histories also allow users to assess the accuracy of data. Specific phrasing of specific subcategory questions is also required in certain cases to address uncertainty. In terms of updating our plan is to update Ordnance Survey footprints every six months and to store demolished building data gathered through this process in our 'Dynamics' section.

As with the Wikipedia and OpenStreetMap model Colouring Cities is designed as  as a low cost model, overseen by expert contributors.  Our landing page also contains a clear statement that data are derived from multiple sources and that accuracy of the data must, ultimately,  be determined by the user. 
 

​

 Colouring London has been designed as a free knowledge exchange platform that collates, collects and generates open data on London's building stock, able to be used by everyone. We do not sell data and we will not share user’s personal data (e.g email address) with any other organisation.
 

The site is explicit in the user agreement, required to be accepted on our sign up page, on the way that contributed data can be used. Colouring London contributions are licensed under the Open Data Commons Open Database License (ODbL) by Colouring London contributors.

Users are free to copy, distribute, transmit and adapt our data, as long as they credit Colouring London and its contributors. If users alter or build on Colouring London data, they may distribute the result only under the same licence.

​

The sign up agreement emphasises that when you make a contribution to Colouring London, you are creating a permanent, public record of all data added, removed, or changed by you as noted above. It is also explicitly stated that Colouring London is unable to accept any data derived from copyright or restricted sources, other than as covered by fair use. Data sources are encouraged to be recorded wherever possible. 

Our platform code are also open and we encourage its use by other cities and towns. Code are available on our GitHub site 

https://github.com/colouring-london/colouring-london  under the following licensing terms: 

'Colouring London Copyright (C) 2018 Tom Russell and Colouring London contributors'. This program is free software: you can redistribute it and/or

modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

​

​

We closely monitor changes to UK policy in relation to the opening up of mapping data at building/property level particularly OS Mastermap footprints,  Valuation Office Agency property tax databases (which contain comprehensive building attribute data relating to many of our categories) and OS historical map data. Our work also involves demonstrating to government why open release of these datasets - with footprints and property tax data now freely available in many other countries- is so important. Once OSMM footprint geometry has been released, all our data will be able to be mapped by all users, as on our site, rather than point data only.   
 
Our data are also specifically designed to support the UK's net-zero greenhouse gas targets for 2050, the United Nation's 17 Sustainable Development Goals , and its New Urban Agenda. Changes to The UK National Planning Policy Framework, and to energy legislation and  UK open spatial data policies are also particularly relevant to our research programme.
 

​

Open Data Commons Open Database License (ODbL) Colouring London contributors are free to copy, distribute, transmit and adapt our data, as long as they credit Colouring London and its contributors. If users alter or build on Colouring London data, they may distribute the result only under the same licence

​

 Data privacy and data ethics are of the highest priority to the Colouring Cities Research Programme.  We work to the best of our ability to ensure we do not contravene any existing ethical frameworks. Each data category is widely consulted on and rigorously checked for potential issues prior to release. Issues on  contravention can be raised by users on our Githib site which is constantly monitored. We also actively discourage the contribution of personal data, avoid the collection of data within the building fabric,  incorporate controlled dropdown menus (with an internal moderation system for sources), and moderate all bulk uploads. We also  state that all data uploaded must be from an open source or generated by the user themselves. Our ongoing work with internal data ethics working groups at the Alan Turing Institute and with national and local partners ensure diverse feedback routes and we actively seek to learn from, and collaborate with organisations advancing the data ethics agenda such as the ODI.  

We are collating and collecting open data on London's building stock to provide  essential information for citizens, researchers, education providers and policy makers, to support the development of sustainable and resilient buildings stocks. . We also want to assist those designing, constructing, caring for, managing and studying London's buildings to help solve urban problems both providing data, and through interdisciplinary/collaborative work .
 
Our aim is to create a one-stop-shop for open data on London's stock. The release of these data is also designed to stimulate the production of innovative and efficient products  within the academic, non-profit and commercial sectors which promote and support the UK's transition to a low carbon economy, the UN's Sustainable Development Goals and the UN New Urban Agenda.
 
We also believe that it is healthy for Colouring Cities platforms collating data on the building stock to be curated by research institutions, whose stance is impartial and whose brief is to undertake research on the built environment for the public good. At the Alan Turing Institute we work within the Urban Analytics programme. 

 

We are already aware of areas of research, such as energy, where demand for accurate building level attribute data is very high. We also know, from extensive consultation, that these building attribute data are also important to the construction and property industry, housing suppliers, planning bodies and the education sector. We are therefore excited about the many ways in which the data might be used.
 
We are currently developing a curated data showcase facility to allow users to upload information, images and links to how data from Colouring London are being applied to urban problems, and to in doing so to inspire and inform.
 

As noted above, building attribute data for London at building level is provided free for all those involved in the design, research, construction, management and maintenance and analysis of London's buildings, and its sustainable development. Our project is also designed to encourage use and knowledge sharing by diverse audiences. This element is central to its design.
 
Our task at the moment is to begin to release data for all our c50 subcategories. The second stage of the project will concentrate on improving data quality, and showing how data are/can be used.

The ODI's 10th item on its Data Ethics Canvas addresses the issue of negative project impact. Could the manner in which this data is collected, shared and used cause harm? Or be used to target, profile or prejudice people, unfairly restrict access?  Could people perceive it as harmful?

All spatial data projects that collect information able to be linked to specific addresses need to be very careful with regard to the type of data collected and how it is held and accessed. A number of checks have had to be put in place to ensure the safety and privacy of building occupants, and platform users. 
 
Examples of ways in which we are working to minimise negative impacts include
a) discouraging the submission of personal data (e.g. email addressed, real names), b) not collecting data on the insides of homes, c) avoiding  freetext wherever possible and using preset dropdowns, to prevent cyberbullying and security risks for occupants, d) only allowing users one vote per user on 'like me?', e) having no negative option for 'Like me?' again to prevent cyberbulling, f) having a sign up page that provides clear guidelines for responsible and ethical use of the site and g) only allowing the copy and paste tool to be used on one building at a time to deter macilous behaviour, and moderating all bulk uploads. 
 
Owing to concerns raised during consultation with regard to privacy and ownership data, Colouring London also only collects data on buildings where the freehold is held by the state or 3rd sector owners.  
 

Comments can also be made on existing discussion threads.  We are also looking at site improvement forms.

The ODI's 13th Ethics Canvas item addresses considerations of people affected by our project. Are we creating potential risks or issues? How are limitations being communicated to those the data is about, and those impacted by its use? And how are we doing this?
 
Colouring London has been designed from the outset in consultation with representatives from diverse sectors. These are listed on our 'Who's involved page'. Our aim is to work with our partners and their networks to allow possible risks, concerns and improvements to be raised at the earliest possible stage, as well as through our public discussions forum and feedback forms. This will help allow ius to anticipate problems and for these to be addressed and features added, adjusted and/or removed as appropriate.
 

We discuss data ethics issues, on an ongoing basis, with colleagues from Turing's Data Ethics Group. Turing's  'Facilitating responsible participation in data science' Special Interest Group, and with our project partners. Users are also able to raise issues for consideration on our discussion threads. 

Our current data ethics priorities (Last updated August 2021) are to improve our user feedback forms and alert features, to try to address issues relating  to security/privacy in relation to free text boxes; and to identify areas of concern with our Colouring Cities Research Programme Partners.