Using The Internet To Quantitatively Observe The World

Humanity’s collective observation power is booming. Smartphones now sit in billions of people’s pockets, whilst numerous private satellites survey major metropolitan areas several times an hour recording metre-level resolution video.

The ubiquity of these digital observations means that almost every significant event that takes place in the world is being recorded in remarkably high resolution; a bomb going off, a car crash, an aeroplane landing. Due to the unrelenting wave of connectivity, these digital observations are being published on the internet, which is swiftly becoming a data store documenting the majority of the world’s events.

This data store holds the answer to fundamentally important questions for society; what is the relationship between political instability and economic growth? Who is most likely to be the target of a cyber attack? Which region of India has the worst road traffic accident record?

Finding the answers to questions like these is the first step to solving the problems that underpin them. For a human analyst, searching through every news article and inspecting every image on the web looking for instances of protests, cyber attacks or car crashes would take a very long time.

Computers, however, are capable of processing such volumes of information almost instantaneously. Machine intelligence continues to excel at replicating the complicated tasks that take place in the human brain. The latest to be crossed off the list include steering corrections for a car on a highway and real-time lip reading.

The internet is a powerful looking glass for observing the world, and we use machine intelligence to make sense of it

At Cytora, our goal is to provide the most up-to-date, granular and holistic view of the world using the internet as our lens [1].

We use artificial intelligence to replicate the part of a human analyst's brain that can trawl through unstructured data on the internet to understand and record events. We use this capability to give our clients a close-to-the-metal view of the world. For example;

  • Identifying every commercial fire that has happened in the US in real-time and historically so that insurers can continually refine and optimise their models of fire risk

  • Tracking every layoff in the US as and when it happens so that investors have the most up to date picture of economic uptick or slowdown

  • Recording every oil truck explosion in the Niger Delta so that risk managers can understand a market where they previously had no data

In doing so we are enabling businesses to turn once disparate information (in news articles, images and soundbites spread across hundreds of thousands of websites) into intelligence about the world.

Forget the back four pages of The Economist

Social and economic data about the world can be derived using the internet as a bottom-up source of information as opposed to the drip fed top-down information that typically comes from a single organisation. This is useful for two reasons: lag and coverage. Typically socio-economic data coming from governments is lagged by months or years. There are many geographical regions in the world where important information is simply not collected.

With a timely and encompassing model of the world derived from the internet, organisations have the power to tackle problems that were previously unsolvable. Events like cyber attacks can be gathered to allow reinsurers to quantify when data breaches can become catastrophes, whilst regulators can keep an eye on the frequency of drone strikes and their proximity to runways and sensitive facilities. Investors will monitor every brick and mortar retail store opening or closure as and when they happen, providing an always-on barometer for an economy’s performance.

It is thanks to the internet and phones being in every pocket that a vast number of events happening around the world are published online, and it is thanks to artificial intelligence that each of these events can be crystallised to populate a row in a spreadsheet.

Going forward our efforts will be focused on bringing the power of the internet as a looking glass to the industries that are critical to society and rely most heavily on modelling the world.

References:

1. We take an event-centric view of the world, where everything can be described as an event that takes place in space and time. See http://plato.stanford.edu/entries/events/