Predictive Analytics – Game On

Published on 05 Dec, 2019

Predictive analytics, a subset of artificial intelligence (AI), is set to bring about a revolution in the coming decades. It will make forecasting ever-changing phenomenon such as weather conditions or ground motion, for instance, seismic movements, more accurate, helping save lives and reduce the scale of destruction. The possibilities of this technology, expected to have a positive impact on almost every aspect of life – physical, social, economic – are boundless, and we have only scratched the surface.

If you have been crying AI from the top of a building and how it might gobble up jobs and everything as we know it, this article is just for you. It’s time to move on!

Artificial intelligence or AI is just a subset of advanced analytics, which is broad and includes predictive, prescriptive, and AI-driven analytics. It is difficult to soak in everything at the first glance, but it starts making sense when we know analytics have been around for more than 80 years. A slight blur still remains because the objective is human-like intelligible output that can help us move things from vehicles to business decisions.

What tangles our thought process is essentially the fact that computers keep on executing orders flawlessly day after day unlike humans whose signatures vary with each passing day. However, the fact is process execution is driven by programs which cannot be altered without an input. Variations too have to be fed as input to get the output in desired permutations and combinations. That’s not intelligence — because the computer cannot think on its own to vary the output. Human intelligence is something different and AI just tries to mimic it by using precisely defined parameters, nothing more. That’s like a mathematical fraction in case you want to quantify. Being self-aware is something different and would hypothetically be an anomaly.

While Tesla is busy with AI and still far away from operationalizing the definition of an autonomous driving system, we can move on to something interesting in its own right — predictive analytics.

If forecasting is what comes to your mind, let’s say it is just too specific and restricted to results such as weather, production quantity, demand and supply. Arriving at forecasts and conclusions is important and definitely not an easy task, but it is mostly achievable through linear calculations. Predictive analytics on the other hand is vast and involves far too many parameters, most likely independent of each other. Predictive text and movie suggestions in your media streaming app are some of the extremely basic examples, but let’s talk about anomalies. An anomaly in a weather pattern not predictable through the usual prediction models would be a good example. Historical data referenced with local current data together with global climate pattern and seismic data could one day predict a sudden surge in the speed of tidal bores in rivers such as the Seine, Amazon, and the Ganges. It would save countless lives and economy.

Risk mitigation, fraud detection, operations, and marketing campaigns are some of the areas that can reap the immediate benefits of predictive analytics.

The immediate beneficiaries for a decade or so are research organizations that need to predict the market for organizations, both from market research and equity perspectives. It includes research analysts and editors, the essential cogs that need fact, logic, and language in tandem to put meaning in sentences. Predictive analytics can change how research is carried out, how you look at datasets, and how much weightage you give to a particular piece of information. In short, it is nothing short of disruptive for organizations large and small because data science is here to stay and for good.

Predictive analytics more than hypothetically promises to revolutionize any process, industry, and business. It is not limited to geography and other variables because it binds a great number of methods from machine learning to statistics to modeling using historical and existing data to predict the immediate and far-away future. Predictive analytics takes in data that might seem random and varied, but processes it to predict future behavior. The predictions are not restricted to rise and fall in prices, but what a consumer is likely to buy based on past behavior.

That sums up the positive aspects and the immense possibilities ahead.

We are staring at a future when predictive analytics might start looking like esoteric science because of the way it predicts events at government, organization, and even at the personal level. With growth in Big Data and cloud-based movement of analytics, impossibility will have to be redefined every now and then.

…Decades from now, we would like to think of a possibility when a small AI program somewhere in the cloud develops a fancy for predictive analytics. God mode indeed!