X09. London prototype: Testing open bulk upload cleaning - colouring-cities/manual GitHub Wiki

Below are slides showing the results of collaborative study undertaken by The Alan Turing Institute and Historic England in 2020/2021. The study looked at the feasibility of disaggregating of listed building data entries from the National Heritage List for England and the spatial visualisation of entries as individual coloured building footprints/polygons on Colouring London (as opposed point data) . The work was co-ordinated by Dr Gill Grayson and Petra Wade for Historic England, and carried out at Turing by Maciej Ziarkowski (research software engineer) who produced the slides below, and directed by Polly Hudson (CCRP research lead).

The study and subsequent discussions with Historic England concluded that to achieve the most accurate spatial dataset for listed buildings in London (and England) APIs allowing for automated geolocation and updating of listed building data were needed, combined with crowdsourced verification from expert sources to ensure the accuracy of building footprints, coloured as listed, on the map. This combined method was also seen as advantaegous in facilitating greater public engagement in conservation principles and goals.

Summary of conclusions & recommendations

• The automated matching results provide additional information which significantly speeds up the process of disaggregation of NHLE entries;
• Further investment in time in automated checking is worthwhile for the NHLE as additional improvements are still possible to make;
• The fastest most cost effective way of releasing an accurate the dataset for London was proposed to Historic England as the systematically assessment by HE experts borough by borough. This was considered to be uneasible due to constraints on time, and cost;
• Disclaimers are still strongly recommended to ensure those planning alterations to buildings check the NHLE direct;
• Manual checking by Historic England, local authorities, amenity societies or others with specialist knowledge will always be required where a high reliance is paced on map accuracy;
• Manual checking results can then be fed back to help improve matching algorithms;
• The matched data combined with the Colouring London restricted interface allows non-technical listed building experts at Historic England to check and correct data; Once this has been done the dataset can be released onto Colouring London;
• Each data entry can then also be verified using the Colouring London verify button;
• Historic England checks can also be recorded in the edit history and used to increase trust in data and expand HE's role as the principal moderator of listed buildings data.