An interview with Philipp Kats
An interview by Alex Karlinsky
Busy with our frustration, we miss the moment when the amazing future actually becomes our reality. Big Data is one of these miracles. The advent of Big Data has been forecasted by many, yet this phenomenon was not fully understood and recognized even when it became a fundamental part of our lives. We talked to Philipp Kats about how Big Data technologies change our cities and us.
An architect by education, Philipp Kats is also a prominent Big Data specialist. He graduated from Kazan State Architectural University and the Strelka Institute for Architecture, Media and Design. Philipp is a co-founder of the Branch Point project. He works as a data analyst and an editor for special projects at the Rambler&Co Studio for infographics. He also teaches at the Saint Petersburg National Research University of Information Technologies, Mechanics and Optics.
How are Big Data technologies used in architecture and city planning today?
First of all, I would like to say that three years ago, when I studied at Strelka, Big Data technologies had just become available. No one knew about them in Russia. One year later, many Russian professionals were already using the best of them and very soon they outgrew even the most advanced. I think it is a very common situation – new technology is being put on a pedestal, praised, but then fairly quickly skepticism appears.
Speaking about architectural or city planning data analysis, I think, nowadays it is a kind of a compromise between modern technologies and traditional analyses. For example, last year I was helping my friend with an architectural student competition in the USA. The city manager provided GIS1 files with very detailed data: transportation networks, volume of traffic, which areas get flooded every five years, high tax rate blocks, demographics, and even information about where puddles appear each year. The statistical accuracy in the USA is pretty high, and data is easy to work with, so we could use it directly without further checking, simply built our proposal on this information.
Big Data doesn’t give you the answers. It helps you ask the right questions
This is, in my opinion, the best way to work with data, when you can take information as a collection of facts and just create your proposal on its basis. It is amazing how differently people can interpret the very same data set and how different the projects based upon them can be.
Google states that their automatic driverless car system can reduce car accidents rate and, at the same time, help to use fuel as well as road space more effectively /Photo: google.com
How do you use Big Data in your own practice?
For a long time, my colleagues Eduard Haiman, Sasha Boldyreva and I have been working on a Branch Point2 project. We tried to work with parametric architecture and develop it further; we all dreamed about parameters-based design. We hoped to create — on the basis of some tricky code — an algorithm, through iterations of which, some new formal solutions could emerge: not the designs we could make ourselves, but new, totally unexpected and beautiful ones.
We wanted to achieve sort of a Bjarke Ingels effect. His buildings look absolutely strange and crazy, but he can always demonstrate you – in a few simple steps – that this is the optimal and the most practical solution. There is always a good deal of speculation in his demonstrations, but his argumentation works.
At a later stage of the project, we realized that while being not entirely unrealistic, this idea does not seem that appealing anymore. We started doubting this design by algorithms. You probably could keep moving in this direction, but it will be hard to get a plausible result.
Here is an important dialectical issue. Imagine you want to write an algorithm, it always needs very basic formal parameters in the first place. In a complex system such as a building or an urban block, there will be very many parameters and you will always need an initial formal idea – a shape that only a designer can create.
Analyzing Zaha Hadid’s projects, for instance, we always see that an elegant formal concept lies at its core, all digital transformations are based on this initial design. The first idea of the author, the first gesture is fundamental. Any genetic algorithm can calculate which of the several ideas is optimal, but it cannot generate ideas on its own.
So the designer’s will is fundamental. How will the amount of human contribution to the design process change with further development of Big Data analysis methods?
I imagine an analytical machine of some kind – a big and complex quantum computer hidden somewhere in a secret chamber, making in the manner of telepaths and parapsychologists predictions and forecasts for mankind.
I personally think that people will always be part of the process. All this stuff [Big Data analysis methods] is called decision making helper algorithms, and the purpose of these helpers is to find flaws and errors in the process, correct them, and, also, reduce human labor. Man will always be a decision maker and algorithms will provide him with everything, except for the decision itself. You need certain technical knowledge to work in this field, but analytics is art and an algorithm is always like a painting.
Drone equipped with a camera can patrol a given territory independently and transfer images to the information center in real time /photo: Kevin Baird /Flickr.com
But data is always incomplete. How can we rely on Big Data if we know that there will be always errors and gaps in our cloud of information?
It’s true. Analysts are often criticized, for instance, for taking into account only those users who are connected to the Internet. Those who do not have computers are not part of the data we work with. This is true, but frankly speaking, if we know nothing about the problems of one ‘babushka’ who is too shy to use Internet, we can simply ignore her. We have to be cynical to some extend: either this ‘babushka’ changes her attitude or her grandson teaches her how to go online.
Another issue which is always a subject of our discussions is that the technology we have for collecting and storing often cannot ensure the stable flow of information without glitches and errors. We also know that it is absolutely impossible to track everything. I mean, we can track some basic things, but the reasons of certain actions or phenomena often remain unclear. Big Data does not give you the answers. Data collection and analysis help you to ask the right questions.
How does the possibility to ask new questions change our idea of the city?
Eduard Haiman invented a term ‘Plagopolis’. The idea is that the contemporary city — and this trend is clear — becomes more proactive and dynamic. It is an environment in constant flux, which also constantly regulates itself. You can identify certain element and describe it with approximation, but it immediately changes and changes other elements around itself. The city cannot be seen as a mechanism anymore.
Is this idea accepted by Russian city planners?
In Russia, there is no clear understanding of what city planning is. We always draw roads and streets first and believe that everything stays exactly the way we envisaged it forever. At the very best, we start thinking about drawing the streets better – and then maybe it works or may be city dwellers rearrange everything according to their needs themselves.
The city reminds us of a proverbial elephant from the story about blind people. One blind man touches the elephant’s ear, another grabs his trunk, the last one grabs him by the ass
Architects and city planners drive me crazy sometimes with their very stereotypical thinking. They just say: “Pedestrians are better than cars!” or “Creative business will turn industrial parks into paradisiacal islands”. I would like to see some reliable data, some research or at least some basic calculations behind these statements, because these assumptions could be true but could be also very wrong.
How can Big Data help us to understand the city better?
The city is always a proverbial elephant from a story about blind people, who wanted to describe an elephant by touching his different parts. We usually work the same way — someone grabs the elephant’s ass, someone touches an ear, someone grabs a trunk. Each of the blind men thinks that he knows what an elephant looks like.
Big data gives us an overview. As a result, when we focus on one part of the city we know that there are also many other components to it – and all of them together constitute one whole. Yet Big data as such does not give us the understanding of all the complexities. And it will never give that understanding.
I regularly get new data on the city. I can always take a closer look into some specific pieces of data and ask: why? The data itself will not give the answer; it just gives you a good reason to start an exploration, a research, or a new story.
Do these explorations inspired by Big Data turn into real city projects?
There is a method of so called ‘city acupuncture’. You basically search for certain key points – places of energy in the city (it could be a block or a building or even a gap in between buildings), and then in these places make well calculated and well planned changes. The changes could be microscopic and require only a very small budget, but their effect to the city could be enormous. Big Data can help identify these key points.
City acupuncture is still largely a theoretical concept, but we have some examples of small scale spatial solutions with an enormous impaction our lives. For instance, the system of traffic lights, those react on the traffic intensity.
Also today, businesses and industries get robotized, and it creates value too. If now drones start transporting goods, city logistics will benefit immensely – working with data collected by them will be much easier than with real truckers.
Technology which inspires me now, and which I hope will have an impact on architecture, is a new project by Amazon. The idea is that a smart loudspeaker is positioned in the center of a house. It listens to all your questions and gives answers. It resembles Siri7 more or less, but in your home. And this technology would probably change our sense of space more than any algorithm.
GIS data combined with spatial modeling algorithms can help forecast the level of insolation on a chosen territory /photo: Trevor Patt /Flickr.com
In other words, the city will depend more and more on the software?
Exactly. Today, the input/output systems and different city interfaces change a lot institutionally. I think that a cheaper and better taxi service could change my life much more effectively than 90 percent of the decisions made by city planners.
For instance, Über and Yandex.Taxi affect my perception of the city enormously. With all this competition between various taxi services, it turned out that taxi in Moscow could be fast, efficient and relatively cheap — much better than in New York, for instance.
A cheaper and better taxi service can change my life much more effectively than 90% of decision made by city planners
Another service which can benefit a lot from «uberification» is, of course, prostitution. Some clients are shy and don’t use prostitutes’ services because of that. To look for a prostitute in the street could be also dangerous and unpleasant. With a phone it is much easier. Of course, pimps will be out of business, but it is a fair deal, I guess. Someone will definitely do it — and soon — in one of the more liberal countries.
Do you think users will be able to manage Big Data technologies on their own in the future?
The complexity of technology will keep growing exponentially, but we will also learn how to pack it properly. Sleek interfaces make it simple. You do not need to think about what is actually happening. You touch here, press the button there and it works. The more you can hide from a user today – the better. People are now a bit scared of all these incomprehensible processes. There is still no technology as the one we have seen in the film ‘Minority Report’, but I think it describes what is going to happen in the future pretty accurately.
What do you think is the next Big Data challenge?
At the beginning, Big Data was just a fancy concept. It is now slowly but surely falling out of fashion, because the most obvious things have been already done. The next important step is to develop technical algorithms in a very utilitarian way. In five years, the position of a digital analyst (quite well paid and quite boring, I guess) will become a standard position at municipalities, governmental offices and business companies.
And the last problem with Big Data is pretty obvious. There are people who know what they are doing and there are people who do not know how Big Data works but make money on it. The gap between professional technologists who work and experiment in their small world, and people who know how to make real projects for the city always existed. We need to find ways for better communication between analysts and practitioners – this is the key to Big Data success in the future.