Big Data real estate analytics
One of the first tasks assigned to Marcelo Cajias, Associate Director of Research, when he joined PATRIZIA several years ago was to create a presentation describing the residential property market in Munich.
“I logged in to the first databank and clicked 100 times before registering the exercise,” he recalls. “Then I logged out, moved onto the second database, logged in and clicked 100 times. I carried on like this for the next three hours, gathering the necessary data. After that I worked through the data and analysed it over two days before starting work on my presentation.”
The whole process took an interminable four days and left Cajias determined to introduce change. “I said to myself, I can’t do this for the next ten years. The whole process needs to be automatised.”
Since then, PATRIZIA has made huge advances in accessing and analysing huge amounts of data thanks to its Big Data analytics’ solution. This uses cloud-based systems, which merge socioeconomic, financial, geographic and real estate data: the platform compresses unstructured data from over 100 multiple listing systems (MLS) into one giant database.
The next step has been to gain insights by analysing the data using advanced econometric tools and machine-learning methods. Essentially, PATRIZIA can accurately estimate net initial yields (the annual rent divided by the price) for different asset types all across Germany. These estimates are derived from big data and closely replicate those calculated by brokers.
Growing frustrated by clicking into a databank 100 times to generate a report, Marcelo Cajias automated the process and created a Big Data solution for property analysis.
In fact, Cajias is confident that its Big Data analytics’ tool places PATRIZIA one step ahead of other investment managers. “What we’ve created is a tailor-made solution to PATRIZIA’s investment process, namely give me the address and I’ll give you the information”, concluded Cajias.
The approach is unique within the industry. Moreover, it’s fast – the solution can estimate the most important data for estimating and evaluating any asset within just 10 minutes. “Nowadays, we put the address into the cloud and after 10 minutes we gain insights into the entire market in Germany, with a presentation of about 45 slides describing the market, construction levels, liquidity measures, price indices, heatmaps, demographic composition and more essential information necessary for any investment decision”
So far, PATRIZIA has been employing the tool for the German and Dutch residential markets.
PATRIZIA’s one-off approach provides several helpful insights, explains Cajias. Notably, it can precisely predict the time it will take to rent out a building, ranging for example between five days in Munich and several weeks in Dresden.
In addition, the tool enables PATRIZIA to set prices by assessing a property’s characteristics, which range in attractiveness according to individual cities. For example, buying a property with a balcony is important in Berlin but not so relevant in Hamburg.
"Nowadays, we put the address into the cloud and after 10 minutes we gain insights into the entire market in Germany."
Finally, it’s accurate when evaluating investment opportunities – PATRIZIA can forecast rent estimates with an error margin of less than one euro per square meter, continues Cajias. “We’re able to calculate by how much a managed property is under-rented or overrented compared with the market estimate. At the same time, we benchmark our properties with the market so that we can carry out a portfolio analysis of these dwellings and create a strategy of how to manage the portfolio -- whether to buy, sell or hold.”
How can big data change the industry?
Elsewhere within the industry, digitalisation and its related technologies is set to introduce radical, sweeping change. Big data can be employed to help customers, investors and real estate companies alike find an appropriate building with desired parameters. And investors can appraise how profitable the purchase of a building will be. Machine learning capabilities mean a huge amount of data can be analysed and processed far faster to support decisions and create thus value.
With its Big Data solution, PATRIZIA can forecast rent estimates with an error margin of less than one euro per square meter.
In essence, the industry is rediscovering itself in terms of what it can do with data. For example, many data providers claim that social media platforms such as Facebook and Twitter need to be integrated into real estate solutions.
Consolidation is likely to take place in three years’ time after multiple attempts to integrate isolated solutions, predicts Cajias. “That is why we’re in discussion with many PropTechs and FinTechs. We want to find out what will be the best solutions in managing our portfolios.
“The industry is taking its first steps. So far, we’re walking, not running."