Building Safety — so we had a go
and moved 1,000 buildings on line.
The need to radically change the way local authorities, and the industry in general, records built environment information has been clear for a long while. However, it takes a seismic event to shift local government. I’ve often looked up from my desk, only to see glaciers go zipping past the window.
For me, introducing change can feel like attempting to drag something huge and heavy, that moves very slowly.
In the summer of 2019, MHCLG wrote to all local authorities and housing associations, requesting that they collect data on the external wall system of all high rise buildings — cladding on tall buildings.
However, I realised that it just wasn’t viable to identify and survey all those buildings using traditional methods. So I thought I’d have a go at approaching the problem with a bit of service design and technology.
First step was to speak with VU.CITY about their 3D modelling software. As a planning department we’ve been using it for a few years now and they have a tool which visualises rising flood waters. I wondered if it could work in reverse. I asked if they could use the meta-data that underpins their model of London (other cities are available) to ID relevant buildings, i.e. those that poked through an 18 meter plane.
They could and there were 1,224 of ‘em. Step one complete.
Step two was to link these buildings with their ownership data — CCOD (commercial and corporate ownership data). Again, VU.CITY helped here by integrating their dataset with that of HMLR’s land and property data set and their title number and UPRN lookup dataset. UPRNs are the Unique Property Reference Number for each property in the country, and any building they occupy. These can be used as the starting block to creating an all important ‘golden thread’, linking various nodes of disparate data together.
This got us so far, but what about their contact details? Turns out that Companies House has an API. That’s that sorted.
MHCLG had given us a template questionnaire for each building, but it still existed as a physical piece of paper. This is posted to building owners, which they print, completed in pen, scan and return to us as an emailed PDF. Not massively helpful, and often incorrect (see below).
Is that three, four or five materials detailed in the form above? On the following page of the same form, the owners claimed that the materials listed covered 140% of the building. AAGGHHH!
But hang on, we’d fixed half of the problem, by using VU.CITY and UPRNs to identify buildings and their owners. So why not just build a web service that enables owners to provide the required information on line, and contact them all using GOV.UK Notify.
I say ‘just’, but that’s a word too often used.
I worked with Unboxed, over six weeks, to design a prototype which sought to improve:
- the council’s identification of buildings 18 metres and above,
- the sending of self assessment questionnaires to building owners,
- their experience of submitting the requested information
- our experience of collating and reporting the information.
Over these six weeks we learned three things.
Identification of buildings generally uses inconsistent and manual methods, such as ‘checking records’.
To get around this, we worked collaboratively with VU.CITY and our in-house GIS and gazetteer team to build a database.
There is currently much concern around the threshold height and what constitutes ‘in scope’ — is it 18m at the top, 18m at the bottom of the top, where does the bottom start, and is it on a slope? Given that there is a significant possibility that this threshold will be reduced to 11m in 2021, and that we have 16,500 of these buildings in our borough, we should consider, now, how we address this.
The advantage of the VU.CITY method is that all buildings can be quickly and efficiently identified, irrespective of height.
As we saw in the first post, the number of lists and spreadsheets owned by different stakeholders is always increasing. However, the problems with sharing and merging those spreadsheets increases proportionally. This was neatly highlighted by Public Health England and the coronavirus track and trace error.
The team therefore created an online database, which acts as a single source of truth. This doesn’t mean that ‘there can be only one’ source, but it pulls in relevant data automatically and presents it to users in a single location.
APIs enable the real time interoperability between this database and external organisations — such as MHCLG’s DELTA reporting system, the Fire Service, conveyancing solicitors or anyone else.
The crucial advantage of this service is that the data is machine readable, meaning that the exchange of data can be ‘real-time’ and automated.
The questionnaire that was originally sent to building owners was a PDF.
There were obvious failures in this — it might be incorrectly addressed or simply ignored, and the answers could be illegible and inconsistent. So the team took a user centred approach to designing an on-line service. Easier for the building owner to complete, it also provides us with a much more consistent and accurate set of answers.
Creating an online survey minimises human interaction, automates the creation of a single source of truth, and makes it compatible with third party authorities, such as the MHCLG and the HSE.
There are of course further developments which will make this service even better, but first we need to test this prototype and put it to work.
Currently, our service only holds the UPRN, address, land registry and companies house data, together with any self-assessment questionnaire responses. However, the infrastructure now exists to add additional information to each building and export the data.
In part 3 I’ll cover what’s on the horizon, how local government can adapt and what we plan to do next.