Big data, big deal?

By Cléo Folkes, Property Overview Ltd, March 2018

Property, or Real Estate to those outside the UK, is slow, inefficient and conservative as a profession. Low transparency reigns, and data analysis comes through after a number of months – if at all. Big Data can help us see more and what is going on now. So yes, I think Big Data is a big deal. To those of you who are new to the concept of Big Data I’m going to explain first what it is, and then how it is being applied in real estate.

Data being used for Big Data-style projects in property right now are mostly not yet genuinely ‘Big’. There are normally a few words starting with the letter v being bandied around when talking about ‘Big Data’. Engineers at IBM use four of them: volume, velocity, variety and veracity. Some might throw in an extra one in for the sake of it: sometimes ‘value’, sometimes ‘variability’.

Let’s start with the first one: volume. The world is full of data, and growth in the amount of data produced is truly exponential. (tip: invest in data centres!?) There are sensors everywhere, things are being recorded with every order, every website visit, every call, every upload. Currently our living rooms are the most connected within each household, but soon our cars will overtake this. Workplaces are increasingly getting sensors connected to the internet, producing reams of data, for a variety of reasons.

UK Land Registry gave away its data for free recently and you can’t kick their data into shape and analyse it with excel or an access database. As it’s big, right? Well, it’s big, but not in Big Data terms. The volume of data that Big Data is set up to be able to store and process is for far more serious volumes that require some really specialist ‘kit’.

What are serious amounts of data? Consider the amount of data generated by mobile phone use in just 1 day across the globe (with 6 billion mobile phones held by 7 billion earthlings!), the amount of data generated by social media sites, YouTube and the likes, or driverless cars. What if you wanted to analyse big trends, to find out what people worry about or what makes them happy right now, what makes them spend money on your product or invest in certain types of property? What if analysing it helps you sell more effectively to retail customers or sell investments just before a market crash?

The second ‘v’ that Big Data should have is velocity, a posh word for ‘speed’. How fast is new data being generated, and how fast it is collected & analysed? Data being produced but not labelled, categorised and processed is useless, and the larger the time gap until it is processed, the lower its value (hey, is that another ‘v’ I sneaked in there?). Analysis needs to be timely and actionable to be really of value.

For those of us who look at property performance data or lettings & sales deals in the market must wait at least a number of weeks until past the quarterend to find out the recent trends in the property markets. What if we could see things instantly? It would transform the property industry.

Variety focuses on how many types of data is being collected within ‘Big Data’. How many sources of data are there that you combine, how do you get to connect them up, how complex is the data and how dissimilar or ‘disparate’ are they? Can you compare apples with pears? Companies like VTS and Dashflow are using software to combine data into something very powerful.

If we can put data sources together like pieces of a puzzle, the overall picture generated can provide enormous clarity. CBRE has started a service called calibrate that astounds me. It is of use to retailers as well as landlords/investors. They take consumer profiling and analysis of retail spend behaviour to a new level. It analyses things like who makes what the type of purchase where at what time of day and how often. They track GPS signals of people’s mobile devices and thus know where people live and shop, and you can combine this with credit card data so you know what they spend. They know where I Iike to eat out and how often I do so, and how much time and money I spend

each time at the restaurants. They analyse consumer behaviour at each location, for each shop and the shop next door (keen to find out how the competition is doing?). They and other consultancies combine data sources to build up detailed shopping habit pictures instantly, creating great value within retail property. Surprisingly few retailers have opted to use Big Data – yet. But I cannot imagine they won’t. It would be silly not to. Evolve or die…

How about Big Data and ‘veracity’ then? With this the focus is on data quality. Garbage in, garbage out! Especially when using data in turn for machine learning it is essential that data is of good quality. What is it that you want to achieve, so think about what type of data you may collect, how often and from what sources.

Can I be nerdy for a few seconds to think about what quality means? You want data that is accurate, trustworthy, consistent. Is it complete? Timely? For consumer data you want to get a single view: no duplication for the same consumer with different profiles. One true record. And for the number crunchers amongst us you would look at whether it meets minimum standards (‘validity’).

Data is being collected via many sensors within buildings to make sure they are efficiently run and optimally used. And if not, it can be quickly adjusted. Sensors help ensure logistics, retail and office units can be used better. Is one corner in the office no longer used for desk use, then let’s turn it into a more productive meeting space. It can help keep buildings and sites secure. What if data was combined with machine learning? You don’t need a security guard at all entrances to an industrial estate to lift up a beam, but systems can read number plates of trucks automatically and open up the barrier? It can process information on how many trucks come in from where, how long they stay, and Big Data can be used to optimise how and when things are transported or how they are stored.

Investors could also benefit from monitoring the markets: is there nervousness, are we on the cusp of another crash, and if we do crash, what can we do best and how, how do we trade at each point of the day and at each point of the cycle? How (fast) are prices changing?

Big Data still has a long way to go to become ever more efficient. Real Estate needs to generate enough data sources good enough to use. However, as with many things, it is our imagination that will limit the uses for Big Data in the end.