How AI can ease the strain on crematoriums

The Crematorium Society of Great Britain needs help finding sites for new crematoriums. As the UK population grows and the percentage of people opting for cremations over burials rises, demand is putting a strain on existing sites, which carry out as many as 3,000 cremations a year. “Families come in one door, have their 20 minutes, and go out the other door so the next family can come in,” says Andy Williams, head of geospatial at surveyor Carter Jonas. In total, 77% of people who died in 2015 were cremated. That number grew to 85% among those who lived up to 45 minutes from a crematorium – and up to 88% if they lived within half an hour of one. The potential for demand is clear. But finding the best places to build crematoriums is not easy. You need to scour the entire country and find land that fits a list of requirements – and you need to find landowners willing to sell. Taking the tech-less route would involve consulting individual agents, driving around looking for land and landowners – and then repeating that five or six times around the country. It would be time-consuming and not necessarily comprehensive. The solution Create a death surface with geospatial data. First, a heat map of death rates across the UK was built. How? The census output areas from the Office for National Statistics – about 175,000 areas across the UK with similar population sizes – can be combined with death rate statistics to visualise the number of deaths per 500 sq m everywhere in the country. From there requirements can be compiled by seeking answers to a series of questions. How far would people travel to a crematorium? Were travel routes significant? How fast do hearses travel? That information helped them to model catchment areas, predicting where bodies should hypothetically be taken for cremation and the extent of services in any region. A lot of the data needed for these projects is freely available, but few property companies invest in the skills to make the most of it. This was then compared with where crematoriums were sited and how many cremations they carried out, producing a map of highest demand. Using satellite imagery and a geographical information system mapper called ESRI, Williams’s team mapped out greenfield sites matching their requirements. Those were cross-referenced with Land Registry data to find landowners and, with Carter Jonas’s database of contacts and projects, the list was narrowed to those prepared to sell. Williams and his team identified between 70 and 80 potential sites in a fraction of the time it would have taken by conventional means. Williams, previously a chartered geographical officer for the Ministry of Defence, says that, outside retail, property has not harnessed geospatial data in the way that it could. The problem is that this kind of analysis is complex. A lot of the data needed for these projects is freely available, but few property companies invest in the skills to make the most of it. Williams’s team includes two developers who write code to build custom applications for clients. “That is unique in the property industry,” he says. “There’s a lot of people gathering data. We are interested in the application.” In-house development Property companies need to invest in the skills to develop these tools in-house so they can take on projects – such as crematoriums – that lack previously analysed data. It also gives them the freedom to mix and match datasets. Relying on different applications and external sources will probably result in compatibility issues, with fewer options to overlay, manipulate or model all the data. The death surface is relevant only to crematoriums, but the methodology can translate to any use. There is also a scarcity of geographers in property – and in the UK as a whole. “There haven’t been too many people like me in the property industry,” Williams says. There are about 700 chartered geographers, according to the Royal Geographical Society, about 10% of whom work outside the UK. Williams estimates that about 75% of these professionals work in government, and 75% of those work for the MoD. The death surface is relevant only to crematoriums, but the methodology can translate to any use. “Once a person like me gets up to speed with that, the approach is transferable: knowing how to select the right data, how to add it together,” Williams says. It’s possible to find the best location for any asset – whether that is electric vehicle charging points or vineyards – by obtaining the right data, evaluating what will drive people to those assets and plotting that across the UK. Could AI do it for me? Williams has few worries about AI replacing him any time soon: “We are a little bit away from artificial intelligence being able to do that. Artificial intelligence is good at doing repetitive things.” That does mean that if someone were to create a model for how to find the best place to put a corner shop, for example, AI could replicate that. He says: “Potentially, if I have done one or two of them and artificial intelligence is looking at how I have approached it, it could say ‘I’ll do the next one’.” But the methodology relies on combining hard data with lateral thinking. That list of requirements Williams’s team compiled at the start of the death surface project was based on thinking about all the factors that might influence a person’s decision to be cremated. Those factors will vary widely between assets in a way AI would be unable to predict on its own. “There is a skill, as a chartered geographer, to picking the right data. It takes experience, judgement and training,” Williams says. And until AI does take over Williams’s career, property needs to think about investing in those skills. A very tech-enabled bucket list Fascination with death and tech goes far beyond finding the right greenfield sites to end your days on Earth. Here are five other ways to bring kicking the bucket into the 21st century. Enlightenment: The Koukoko-ji temple in Tokyo is home to Ruriden – a mausoleum with more than 2,000 LED-lit Buddha statues. Behind each glowing Buddha is a compartment where people can store their ashes. The lights change depending on the time and season, and visitors can swipe a card to illuminate the statue in front of their loved ones’ ashes. Death insurance: With people and tech growing ever closer, health insurance needs to follow suit. Richard Craib, founder of hedge fund Numerai, tweeted last year that the company offers whole-body cryogenic freezing. A job description posted at the time on Numerai’s website read: “Numerai cares about its employees beyond their legal deaths.” Digital crystal ball: Scientists at Stanford University have developed a program that analyses electronic health records for patients to predict the probability of mortality over the next three to 12 months. Predicting when high-risk patients are likely to die can make it easier for doctors to make death as painless as possible. Cake or death? As long as we resign ourselves to the fact that we won’t live forever perhaps we should plan for that inevitability. Digital health start-up Cake has developed an app where you can plan what happens to you when you die – anything from the type of funeral you want to what happens to your social media accounts – and it gives you options for the services and decisions you should make. Augmented eternity: Can you build a digital version of yourself that outlives you? Can that digital talk to others after you have died? Hossein Rahnama, professor at MIT Media Lab, is working on just that. As people generate more digital information and more online interactions, that data can be used to create a digital avatar. Suddenly, you can become the voice of a chatbot or another Siri on an iPhone. Even after you die, your digital self can live on and nag your loved ones. To send feedback, e-mail karl.tomusk@egi.co.uk or tweet @ktomusk or @estatesgazette