We caught up with Barry Dillon Head of Advanced Analytics at the Aon Centre for Innovation and Analytics @, to get his take on the value of data for predictive analytics. Speaking at the Predict Conference @ September 15 – 17th
How did you first become interested in data / predictive analytics?
I have always enjoyed computers and it seems taking data and predicting things is about the most interesting thing you can do with a computer. It compares well to processing payrolls or doing book-keeping which I guess were the most common computer tasks for many years. World of finance used to offer great opportunities for modelling price movements.
— Aon (@Aon_plc) September 4, 2015
How do you think data will change business (the world or your sector)?
Insurance is a more complex business than generally perceived so the human element and in particular the relationship management element seems unlikely to disappear anytime soon regardless of what happens with the data. The special knowledge of the various experts will become more widely available across the firm but you’ll still need the expert I believe.
What is your best data (data modelling/predictive analysis) tip?
Always assume the data has lots of noise in it no matter what you’ve been told and spend lots of time just looking at the data rather than running analysis and looking at the output.
What was the most surprising thing you have found/uncovered from data analysis?
I was analysing order flow within US bank and one trading desk seemed to have infinite limits. It turned out that their unwinding trades they wrote on paper and had a courier take to Jersey where they entered them in another system that I didn’t have access to. They had been hiding the largest trades this way for years.
— Predict Conference (@predictconf) September 13, 2015
What advice would you give to someone just starting out on their data journey?
Make sure you can do Leaving Cert maths really well as the same old techniques get re-used over and over. Integration maybe less so but differentiation and optimisation just keep coming up and it’s a shame to walk away and say you’ll need an expert for that.
What skills do you think a good data scientist analyst should have?
They must enjoy nibbling away at possibly tedious problems. The sort of people who can happily spend hours untangling a mess of cables or a ball of wool as the data is always messed up and needs untangling and that consumes most of one’s time.
What resources would you recommend (e.g. books, websites, blogs, data, technology, etc.)?
I believe the technology is getting easier and easier to use but people are perhaps getting trickier and trickier to handle. To that end take ability to do things with a computer such as Python as read and perhaps focus on becoming good at communicating your ideas and explaining your solutions. This is where technophiles most often fall down.
Which trends in this area will make the biggest change in people’s lives in the next 2 – 5 years?
The internet has made the world very flat in terms of access to education. How that impacts our attitude and pay for experts is still unknown. In professions such as medicine relatively difficult predictions such as reading retina images to diagnose diabetes will be done better by computers than the doctors and ophthalmologists currently doing the job. That has to be quite a shock to many sophisticated professions. Hopefully the drudgery element of their work is removed and they can take their game up to a higher level. The worst outcome would be to just replace them with computers to save money.
Barry is a mechanical engineer who has worked in finance and industry and has always been involved in finding patterns in data. Barry’s finance work was in Dublin (Davy) and New York (Merrill Lynch) and he also did a Wharton MBA in Finance and Statistics finishing in 2000.
Barry’s remit has always been about turning data into dollars which might have involved bonds, stocks, options or insurance. He has been leading a team of analysts in ACIA for 5 years where they are helping decipher insurance data from all over the world.
Barry Dillon | Head of Advanced Analytics,
Aon Inpoint | Aon Centre for Innovation and Analytics