Real Estate Analyses by Means of Algorithms
Interview with Matthias Standfest, Founder and CEO of Archilyse
Archilyse is a research-driven startup with the goal of comprehensively measuring architectural quality. Archilyse uses simulations to generate a holistic analysis for real estate decision-makers. Archilyse is the only PropTech globally which not only takes the geographical location into account, but also analyses the indoor floor plan qualities and the environment. The data is then displayed in easy-to-read heat maps or can be exported in the form of key figures for further processing. Until now, project developers, architects, planners and property owners have used this analytical data to create added value and save costs at every stage of the value chain.
In 2017, you founded Archilyse. What is your motivation, your vision? Why did you start this company?
A question that has been following me around since the beginning of my studies is: “What is good architecture?”. What are important quality factors for architecture? In my research at ETH Zurich, I have investigated how human biofeedback correlates with spatial perception. We measured the heart rate of test persons in specific urban settings and deduced that certain architecture generates stress. This research, and finally my PhD, was in the areas of machine learning, geometry, and human space perception.
When I got in touch with the Swiss real estate industry during my doctorate, I realised that massive data potentials were lying idle. I wanted to build on my research results and make them applicable to the practice with the help of new technologies.
In my doctoral thesis, I was able to show that it is possible to automatically retrieve data and qualitatively evaluate architectural features from floor plans or 3D models. This brings a semantic level into the evaluation of architecture that was previously not possible. With this method new tools for valuation, planning, marketing and consulting in the real estate industry can be established. Risks or opportunities, for example of planned buildings, become apparent more quickly and with little effort — and architecture becomes better in the long term. That drove me to found Archilyse.
What do you offer your customers?
The secret of Archilyse is to use data and information with the help of new technologies in such a way that it is very easy to achieve added value. Using algorithms, we can generate extensive analytical results from floor plans or 3D models. Normally, we get a 2D floor plan with the appropriate address, convert it into a 3D model, and calculate for every square centimeter how the 3D view is, e.g. How many percent of my view allows me to see the lake or a particular landmark? How high is the street noise that affects the apartment or how is the sun radiation in December?
The novelty is that we also include the neighborhood in these analyses: do buildings block my view or do trees swallow sound?
So far, neither the location nor the floor plan have been evaluated in such detail. Factors such as the furnishability of floor plans, the view, the neighborhood, the infrastructural connection or the distance to the nearest supermarket usually are ignored. Our USP is that we can incorporate numerous and complex features, such as those already mentioned, into our analyses and thus make statements about how well people feel in a specific architecture.
By translating current scientific knowledge into algorithms and including them in our process, we can also consider factors based on the latest research findings. Our results are already being used by leading companies in the Swiss real estate market.
What is the added value of your analyses?
Our analyses are equally useful to portfolio managers, real estate developers and appraisers, as well as real estate platforms and advisors:
For portfolio managers and valuers, the generation of various additional information in the form of data on their real estate is interesting. We provide them with view, sun simulations and noise forecasts as well as data on the quality of the floor plan, distance or travel time to urban sites. This goes far beyond the features of the classic rating or evaluation. This information helps to make the rental price estimates more accurate and thus also to minimize vacancy.
For planners and developers, we offer automated norm checks and automated controls of the construction zone regulations. We show with our software whether e.g. the SIA norm 500 complies and the building is thus barrier-free. Or we determine the maximum possible construction volume and thus ensure greater planning security.
For real estate platforms, we also offer address-based values that do not require floor plan information. And for office planners and consultants, we have developed a workplace analysis and optimization software. With our automated office test planning, area and performance analyses as well as facility analyses and comparisons can be carried out according to customer-specific requirements. This makes it quick and easy to optimise office planning and furnishing.
Our results are all available in real-time, meaning, for example, that as a planner, I can continuously monitor the effects of my design process, making it easier for me to come up with an optimised plan more quickly over several design iterations.
Who benefits from it? What are the effects if the subjective, human assessment is replaces by a virtual one?
Currently, some apartments are being offered too expensive, others are too cheap. We help to create a more balanced and transparent interface between rental needs and availability. This also helps that the market is balanced out.
Is it also possible to use your data for other purposes, e.g. for the selection or optimisation of production sites and plants?
When selecting and optimising production, our software can support in a simple and uncomplicated way. On the one hand, we can facilitate the selection process of a new location by incorporating topographical and infrastructure-specific aspects and requirements into our analyses. In doing so, we consider factors such as the infrastructure, the view, noise pollution, travel times to customer-specific destinations and solar radiation. These functional qualities can be integrated as parameters and used to determine the best location. On the other hand, existing production sites or plants can also be optimised. We can check whether the routes are optimally planned and the escape routes are efficiently organised.
Who owns the data and how safe is your system? What third party data do you use?
We only use Swiss server providers for our Swiss customers. We have our own data rooms for each customer. Basically, the data we work with is less security-relevant — it is more the question what our customers do with our results.
Can you tell us about specific examples of where you could add value with your product?
Although we already have some very successful applications, I definitely have a favorite project. With one of the largest Swiss real estate portfolios, we were able to work out an overall solution. We pick up the plan boxes of the archive, scan everything and bring the boxes back to the archive. At the same time, we then process all digitally available plans, automatically generate IFC models, simulate all parameters and even provide a rental price valuation through a partner company — optimised for each individual apartment. This not only results in higher rental income, it also drastically reduces running costs. This project shows the potential of our application and how property decisions will be made in the future.
Where do you see the development of your company for the future?
In profane and repetitive tasks, such as the design of residential buildings and office buildings, where there is a high pressure of innovation, algorithms will certainly intervene massively. Some activities won’t be continued in the future if, e.g. it is no longer necessary to check the floor plan for compliance with SIA norms, because the computer has already done this before.
We are at the crossroads of a future that will be dominated by climate change and digitisation. We have an obligation to be responsible and to develop the best tools for the tasks ahead.
The time that is being saved by using our method can then be used to create something new and to develop better, creative solutions.
Do you want to know more about this? Check out our articles about Architecture Analysis on Medium or follow us on Twitter.
