What is Predictive Analytics?

Predictive Analytics is a term being increasingly used in sectors from healthcare to marketing but few people understand what it means or how it works. Predictive analytics refers to the use of statistics and modeling techniques to make predictions about future outcomes and performance by looking at current and historical trends in data to determine the probability of potential scenarios occurring in the future. For example, past records and present data of a region's weather condition can give meteorologists information to forecast future weather. By applying the concept of prediction, there is a fundamental shift in focus from ‘descriptive analytics’ – understanding and articulating what happened in the past and why – to using the available data to anticipate what will happen in the future. This powerful set of statistical tools provides organisations with the advantage of determining potential outcomes and using this information to plan action and allocate resources preemptively.

Evolution of Predictive Analytics

Predicting the future has been a constant preoccupation of humans. Throughout the millennia we have developed a variety of tools, heuristics and superstitions designed to create the feeling of control of things we cannot anticipate or comprehend and the confidence to act. These tools have evolved over time to become more sophisticated and objective by relying on empirical data.

The early statisticians used pen and paper to draw insights from historical data, with one of the earliest adopters being Lloyd’s of London in the 17^th century who utilised a predictive model to calculate the risk of ship voyages from records of past journeys to ensure profitable investment. Fastforward to the present day and an increasing digital world, these predictions are no longer made with pen and paper and unreliable data in handwritten ledgers, but huge, vast and complex data sets that are growing exponentially and continuously curated to fit specific needs.

Laptop showing Data Analysis — Nowadays, we use huge, vast and complex data sets.

AI can use these data sets to generate predictions across any number of contexts, from what product to suggest someone on Amazon to identfying pedestrians crossing the road with self-driving cars.

An important thing to realise is that every prediction is based on a probability; so even when a prediction fails to transpire, AI uses this ‘failure’ to learn from the experience and inform future judgements and decisions. In the context of AI and machine learning, prediction is all about data.

The rise of AI

In the 1950s, the concept of Artificial Intelligence (AI) was introduced to the world as the science and engineering of making intelligent machines by mimicking the problem-solving and decision-making capabilities of the humans to perform complex tasks. AI does this by analysing massive amounts of historical data to “learn” patterns in the data set and generalising these learnings to new data, enabling ‘intelligent’ automated action. Statisticians realised AI had the potential to push predictive analytics to another level and overcome much of the tedious and tiring search for patterns within data. AI is a tool that naturally exceled at this task and would do much of the work at a speed and accuracy humans are incapable of. Humans would only need to prepare the data, calibrate the algorithm and examine the results.

Examples of Predictive An