Just had a great discussion about data analytics and visualisation at the SWIFT Nordics Conference in Oslo. During the discussion we talked a lot about data for risk management and how real-time tracking of payments could help identify suspicious activities from fraud and malware attacks to the possibility of AML and sanctions breach. I couldn’t help but think, during the debate that I’ve been going at this data challenge for a long time now.
In the 1990s, we talked a lot about predictive analytics, propensity modelling, 1:1 marketing, cross-sell, share of wallet, single view of the customer and such like. Two decades later, we are still talking about the same thing, although the tools are changing. The tools today are networked tools tracking digital footprints; the tools twenty years ago were massive data warehouses to crunch data and identify themes. The tools today are data visualisation for spotting unknown unknowns; the tools twenty years ago were lucky if they could find a known known.
Anyways, the big thing that came out of the chat for me was that in the 1990s, I was involved in a firm that offered neural networked trading systems. The idea was simply to align algorithms so that if one stock price or currency or corporate action took place, the algorithms would look for linked activities and pair them. So, for example, if IBM share price improved combined with a lowering of Microsoft’s share price, the system would sell IBM and buy Microsoft. If the movement went the other way, then the algorithm reversed.
Simple stuff, but that became the backbone of blackbox algo trading in the 2000s that led to high frequency trading and the Big Short years later.
So then the big thing of the 2000s was the semantic web. The web would learn all about you and would proactively predict what you needed. It would build context and relationships around you, and would be there for everything you need. Rather than ask questions, the semantic web would just give answers. Beautiful. It’s not here yet – you need this decade’s technologies for the semantic web, more in a minute – but what did come to fruition is the social web that provides semantic capabilities. The social web is illustrated by tweets about stuff that could be important – company activities for example – and is now linked into the algorithmic trading systems through feeds such as StockTwits. Auto sell and buy based upon social media discussions is here and now. Equally, the semantic web is illustrated today by Facebook apps that tell me which of my friends are near me now, sure I could use Foursquare or others for this but hey, I’m old, so I used Facebook.
So then we come to today’s Big Data thing and it’s all about Deep Learning. I blogged about this recently, covering the activities of Google, Facebook, IBM, Microsoft and more. Deep Learning is where the machine becomes more intelligent that the human, playing video games against itself and never losing and clearly beating mere muggles at Jeopardy! Great. Combine this with neural networking and the semantic web, and you have the dream vision where Ultra High Net Worth clients of Wealth Management companies can become even more Ultra by just following their deeply learned robo advisor. Fantastic.
But what comes after that?
That was my final question to the panel today. They talked about machines trading with machines and no payment information is ever seen. The payment transactions are just managed on the network, visualised in real-time dashboards and analysed to ensure you never knowingly overspend.
But I have a different vision. You see scientists have just had a breakthrough, and found the part of the brain that determines evil. As The Telegraph puts it:
The ‘root of all evil’ has been discovered by scientists who found that a distinct part of the brain fires up before premeditated nefarious acts are carried out … [Researchers say that] it might even be possible to continually monitor brain activity and alert health experts or the security services before an aggressive attack, such as in Philip K Dick’s dystopian novella Minority Report.
Shoot. Don’t tell the Hulk!