## <span class='subtitle notbold'>Sino Townhall / Oct. 2023</span><br /><br><br> ## Real Estate<span class='subtitle'><br>And its Impact on Society<span> <br> ## <span class='subtitle notbold'>Professor Thies Lindenthal</span><br /> <br> <br> <img src='imgs/logo.png' width=370>
<br> <br> ## <mark>3 thoughts</mark><span class="subtitle"><br/></span> <br> * <span class="transparent70"> 1. Real Estate is as diverse, multi-dimensional and interconnected as the societies it serves.</mark> * <span class="transparent70"> 2. To assess and optimise, we need to quantify first.</mark> * <span class="transparent70"> 3. Over-simplification of complex systems is dangerous.</mark>
<span class='transparent'> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <span class='source transparent30'>Photo: <a href="https://www.flickr.com/photos/alexbrn/">Alex Brown</a></span> </span>
<span class='transparent'> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <br /> <span class='source'>Photo: <a href="https://www.flickr.com/photos/seier/">Seier</a></span> </span>
## <mark>people and environment</mark><br><span class="subtitle white"><br/>Impact on others growing concern: How to create green and healthy spaces?</span> * <span class="transparent70"> Tenants expect sustainable buildings (Fuerst et al.)</span> * <span class="transparent70"> Health not an afterthought anymore.</span> * <span class="transparent70"> Quality can still compete with home office.</span> <br> <br>
## <mark>Quantification</mark><br><span class="subtitle"><br/><span class="white">Measuring economic activity inside buildings easier than ever.</span></span> * <span class="white"> Tracking of shoppers inside Beijing mall based on CCTV, linked with till data (Jiabin Wei, 2022)</span> * <span class="white"> Modelling of synergies and optimisation of tenant mix</span> <br><br><br><br> <span class='source transparent50'>Photo: <a href="https://www.flickr.com/photos/musicsthename/">Ava R </a></span>
<h2>Assessing Potential<span class="subtitle"><br />Retail vacancy data is scarce in the UK: So we analysed high street windows.<br /></span></h2> <p class='source'><img src="imgs/mphil-diss-emma.png"> Source: Emma Waterhouse “Applying Machine Learning and Image Classification to the Built Environment: The Case of UK High Streets”. Photos: Google Street View</p> * Which neighbourhoods of the UK are flourishing: Economic prosperity, influx of capital and people, tourism? * Where are high streets in decline?
Our Toolbox
Computer vision + Machine Learning (ML) + trad. econometrics
## <mark>Quantifying style(s)</mark> <span class="subtitle white"><br><br>“Building Better, Building Beautiful”? </span> * <span class='transparent70'>The UK Government has been emphasising the "beauty" of homes and neighborhoods: _"We should then build as our Georgian and Victorian forebears built [...] All objections to new building would slip away in the sheer relief of the public"_.</span> * <span class="transparent70"> Not supported by transaction data <a href="https://link.springer.com/article/10.1007/s11146-021-09845-1">(Lindenthal & Johnson, 2021)</a></span> <p> </p> <p> </p> <p> </p> <p> </p> <p> </p>
## <mark>Diversity</mark><br><span class="subtitle white"><br/>Tenants' needs and preferences differ!</span> * <span class='transparent70'>If you look close enough, no one is the same.</span> * <span class='transparent70'>One size fits all? No.</span>
## <span class="transparent70">One taste for all?</span><span class=" transparent70 subtitle"><br/>How diverse are aesthetic preferences? (Lindenthal, Hoeltken & Wang)</span> * <span class="transparent70">App available at https://4walls.cremll.com</span>
## Some general trends...<span class="subtitle"><br/>People like greenery, space, good quality—and often historic buildings.</span> <div class="figure"> <img src="imgs/example_like.png" height="500"> </div>
## Better not...<span class="subtitle"><br/>Neglected, cheap materials, mixed styles.</span> <div class="figure"> <img src="imgs/example_dislike.png" height="500"> </div>
## The Marmite homes<span class="subtitle"><br/>The really interesting cases: People don't agree. Taste, largely, remains personal.</span> <div class="figure"> <img src="imgs/example_mixed.png" height="500"> </div>
## <mark> No over-simplification </mark> <span class="subtitle"><br/><br/><span class="transparent70"> Stronger analysis with bigger and multi-dimensional data</span></span>
## From utility to investment returns<span class="subtitle"><br/>Value is created along a long chain: people → companies → buildings → portfolios.</span> * Traditonal approaches have to simplify—much information is lost. <p class="figure"> <img src="imgs/value-chain.png" height="450"> </p>
## Predicting investment returns<span class="subtitle"><br/>ML finds predictability where traditional approaches fail (Leow & Lindenthal, 2023).</span> * REITs are more predictable than other stocks. Implications for multi-asset portfolio allocations. <p class="figure"> <img src="imgs/value-chain-step1.svg" height="450"> </p>
## Deeper analysis of performance<span class="subtitle"><br/> Detailed building data w/o aggregation improves benchmarking at company level.</span> * Stars and hogs cannot hide in crude aggregates. More details and interactions considered. <p class="figure"> <img src="imgs/value-chain-step2.svg" height="450"> </p>
## Full chain: Tenant → investor<span class="subtitle"><br/>Benchmarking and optimising without loss of detail. Many KPIs possible.</span> * Where is value created? Learn from **ALL** data to improve health, sustainability and bottom line. <p class="figure"> <img src="imgs/value-chain-step3.svg" height="450"> </p>
## <mark>Summing up</mark><span class="subtitle white"><br><br/>To empirically test big questions we need big data</span> * <span class="transparent70">Real Estate is diverse and multi-dimensional. Buildings never stand in isolation but are closely connected with communities, the environment, and the economy. These links matter!</span> * <span class="transparent70">New data and machine learning allow us to quantify the wealth of ties between the built environment and society.</span>
