Highlights from Predict 2016 — The Data Science Conference


It is an exciting time to be in data science where your ideas and innovations can impact on millions of potential customers globally. Disruption on a global scale is now open to us like never before.

And yet success in data science is not as easy as you would think; it is an uncertain time, not least because of data and the new digital economy making their presence felt across many sectors.

Organisations are using data to compete and thrive in the new economy. A chasm has formed between the digital haves and have-nots with some established companies struggling to adapt.

Predict was created to fill a void for an outward-looking, data science conference that brings together leading international thinkers, practitioners and decision makers in data science, government and business.

I am excited about how we as a community, can rise above the noise to deliver world class data science and analytics for customers globally, and to use data and analytics to compete and to make an impact at global scale.

We like to say in Creme Global that “Data Science is a Team Sport”. We believe that organisations can’t do it alone. They need to engage with the community, in partnerships to deliver their data science projects.

That is why we created Predict; to build the ecosystem and to create a network that will inspire people, cross-pollinate ideas and demystify data science for all sectors.

Something special is happening in Ireland, with global companies setting up innovation and data science headquarters here. If we look at the Predict partners: Aon, Accenture, Zalando, CeADAR and others — each is innovating and investing heavily in this area.

Highlights

One of the interesting things at Predict this year, was the extent to which the technology was almost being taken for granted. Instead of focussing on the technology, the softer skills and crucial elements like culture, teamwork, collaboration and trust were regularly being highlighted and discussed.

John Elder pointed out that finding the answer and proving it, doesn’t mean it will be used. He highlighted why data scientists must build trust as well as great models, so your work doesn’t end up on the shelf.

Shaku Atre also talked about the importance of storytelling, a three act structure, to make sure clients get the point of the data gathering exercise with a beginning, middle and end. We heard about emotional rather than rational engagement with people; and the difference between being a marketable business as opposed to marketing a business.

Bernard Marr explained that everything starts with strategy and offered us three strategic lenses on data strategy:

  1. Use data to inform decision making
  2. Use data to improve operations
  3. Treat data as an asset, something to be monetised and make the business more valuable.

We got a deeper dive into the technology piece with Todd Curry from Aon: Crunching big data demands new technologies. He walked us through:

  • Cloud vs Laptops and the benefits of moving to the cloud
  • Change Data Control vs ETL
  • Ephemeral vs. persistent computing — again a benefit of using the cloud, and
  • Agile and scrum versus waterfall development

Karen Church from Intercom gave great insights into building a data analytics capability inside an organisation through a hybrid combination of centralised and embedded teams.

And as with last year — there was a lot of talk about disruption.

On day two Sean O’Sullivan told us about some of the SOS Ventures investee startups that are augmenting humans’ ability to become and stay healthy. He explained that devices that close the feedback loop between measurement and habits enable behaviour change that can have a huge impact on health.

Dermot Casey offered some real food for thought, reminding us that our models are formed opinions created in code, mostly created by white, middle class males. This results in a biased reflection of the world. We used to worry about the input data and “garbage in, garbage out”, now we need to be more concerned about “bias in and bias out”.

New this year was a Design session in which Robert Coyle chaired a set of talks on design. In his work in design for artificial intelligence, Padraig Mannion from IBM explained that people are naturally pre-disposed to interact with technology in a very social and natural way. This can be used to a designer’s advantage, but he also warned that people will be very intolerant if that technology doesn’t deliver.

We learnt a lot about how to ensure success for our data analytics projects. Uli Bethke from Sonra gave us three elements that are worth committing to memory: context, vision and outcomes. Aoife D’Arcy also emphasised the need to articulate a clear vision (using the architect Dermot Bannon as an exemplar) for your data analytics project and to engage with the business to explain it and understand the decision space.

We had a session on Food, Health and the Environment — chaired by Cian O’Mahony — and we were dismayed as Richard Corbridge showed a living room sized office full of stacked cardboard boxes, each with dozens of paper files, which hold the medical records of yesterday’s, today’s and tomorrow’s admissions in an Irish hospital. All but one Irish hospital uses paper rather than digital records. The good news is that the government has committed to a significant investment to digitise all of our hospitals.

Another new session this year was on Open Data with Dennis Jennings chairing a discussion on how we can get more value from the data that governments create and manage on behalf of citizens.

Finally, the last takeaway, based on what we think was the most used phrase of the event — something that was one of the reasons why we set up this conference in the first place.

And that was of course: Data science is a team sport.


About Predict — The Data Science Conference

Over 580 professionals in data and analytics attended the second Predict Conference held in the RDS, in Dublin in October 2016. Run over two full days with presentations, panel discussions and exhibitions — and with two additional days of hands-on workshops — the Predict conference brought together the key players in business, government and research to share learnings and insights and to discuss what data modelling and predictive analytics means for decision making in the data age.

The Predict Conference will publish an eBook on proceedings and will provide access to presentation slides and videos to attendees. We also have some amazing video and presentation content coming up, and of course — Predict will return in 2017!

I look forward to continue to work with all of you — through the Predict Conference and other initiatives — to help organisations to develop world class capabilities in Data Science.

I would like to say a big thank you to all of our attendees, speakers, partners, exhibitors, volunteers, suppliers (including Avtech, With Taste and the RDS) and not least to the Creme Global team, without whom Predict could not have taken place.

Follow the Predict Conference on Twitter: @predictconf or sign up to the Predict Conference newsletter for updates.