Swiss start-up PriceHubble is one of the innovators in the field of Automated Valuation Modelling. It uses machine learning and big data analytics to provide valuations, value predictions and data visualisations to companies across the real estate industry.
Loeiz Bourdic, Head of France for PriceHubble, explains what it brings to the real estate market and gives his views on the current state of development of Automated Valuation Models (AVMs) and predictive analytics tools for real estate in Europe.
We use machine learning and big data analytics techniques to improve transparency in real estate markets.
First, we collect and aggregate data from open source or commercially available datasets, data sharing agreements or exclusive data available to us through partnerships. For both the residential and commercial real estate markets, we focus on historical data and forecasts on real estate transactions and offers. We also analyse vacancy rates, office take up, and the corporate tenants in each commercial building.
In order to enrich our analysis and predictions, we overlay urban data such as amenities and building characteristics; socio-economic data, such as income distribution, unemployment rates, types of jobs and companies present in a neighbourhood; transport data; and environment evolutions, which refers to factors like new transit infrastructure and gentrification dynamics.
We then run multiple analyses on these datasets to extract relevant and valuable insights. We have developed a standardised, process-based approach for each dataset. The first stage involves analysing past and current trends and patterns, to quantify the impact of local characteristics such as socio-economics, infrastructure and urban amenities on real estate dynamics. Then, we use the patterns to forecast the impact of urban evolutions on real estate markets and present the insights through data visualisations.
The tech addresses fundamental challenges that most of real estate players are facing. These include getting access to up-to-date and reliable data to better inform decision-making as well as being able to value any real estate object in a reliable, accurate and fast way. It can also help those in the real estate industry to anticipate market dynamics and the impact of urban evolutions on their assets.
There are many players developing AVMs for the real estate industry, but we consider these models to only address one part of the problem. We are going much further by leveraging big data and machine learning to anticipate evolutions in property value and market dynamics.
Our offers are mostly co-developed with clients, whether they’re banks, credit and financial intermediaries or insurance firms wanting to better advise their customers in Switzerland or Germany or institutional investors, developers and asset managers in France, looking to implement new investment strategies, analyse investment opportunities and improve portfolio management. The insights help them invest, advise and close deals faster.
The quality of valuations and forecasts relies on the quality of the data used. We work hard to get the most detailed, reliable and exhaustive data sources we can from our providers, from our partners, or by crawling publicly available information on the residential and office real estate markets.
We’re very open about what data we have and what we don’t have. We believe being transparent with clients can only increase their trust in our solutions and generate discussions on the new data we can integrate.
In France, AVMs are yet to be widely used as Credit Risk Management has historically been based mostly on household solvency rather than on the value of the real estate property. However, recent legal changes have opened up new opportunities for AVM players.
Meanwhile, in countries like Switzerland and the UK, AVMs have long been used and the challenge is to take the model further and leverage forecasts and predictive analytics tools.
The real estate market is becoming more transparent. In order to generate better alpha return, market participants need to be more informed to make faster and smarter decisions. Funds, investors, and asset managers are competing to gain access to the most reliable and up-to-date information. With data being more readily available, the biggest opportunity for automated predictive tools is to assist players in making better investment decisions with accurate analyses, valuations and forecasts.
Date of creation: 2016
Key investors: individuals and corporate investors, including SwissLife & Helvetia
Locations: Switzerland, France, and Germany
Markets addressed: Residential and Commercial Real Estate