image Bath Digital Festival 2021

Synopsis

We are awash in a sea of data - how can we make sense of it all, and turn it to our advantage? In particular, how can we exploit data to deliver complex change? Machine learning / AI / algorithmic bias - these are all 21st century terms.

But what do they mean, and how should we adapt and evolve? How can modern business exploit the exponential growth in data science to deliver projects and business change more effectively?

In this talk James Lea, founder of Project Science, shows how 21st century industry is adapting to these challenges, and how the best are exploiting data to deliver complex change more effectively than ever.

Here are some of the key points made.

image Bath Digital Festival 2021

The Big Questions

Central to safe, timely and cost-effective delivery is the art and science of making predictions.

  • What is data?
  • How can we make predictions using data?
  • How can we exploit data to deliver complex change?

Building prediction models

How do prediction models connect the past with the future?

  • We establish a relationship between things we currently observe, and their likely correlation with future outcomes: a type of model
  • Use the past as a guide to the future 2 : train the model.
  • The better our past knowledge, the greater our prediction capability

However, beware the Black Swan (Taleb), or the fat tail: rare events are perhaps not so rare, and throw our models (and predictions) out.

image Bath Digital Festival 2021

Types of prediction model

We routinely use prediction models every day in our lives. The models we use (should) depend on the consequences of getting it wrong.

  • Guess
  • Intuition / gut feel
  • Pattern recognition based on experience
  • Physics - core representation of problem
  • Ensemble (multiple techniques, multiple views)

What if your money or your life depended on getting the correct answer?

The 21st century will bring substantial change:

  • Ethical data sharing drives learning from experience
  • Citizens take control of their data
  • Better models of the work: process mining and discovery - exploiting the shared data sets (probability-based)
  • Data-driven delivery culture: clients will expect more
  • Expect disruption: Automation; Metaverse; Quantum computing