Perceptions and popularity of analytics technologies over time

Will machine learning be the dominant technology focus over the next 2 years?

While doing some market analysis, we decided to take a look at how search terms were being used in Google around the world as they related to advanced analytics technologies. We picked five terms to compare over five years: risk analysis, big data, machine learning, Monte Carlo method, and forecasting. We then proceeded to download the data and apply some quick and dirty forecasting to see what would happen to the popularity of the search terms overtime.

Quick Analysis of interest in analytics technologies over time

There are two very interesting things to note in the chart. First is the spike in relative search popularity of forecasting which occurred during the last 2 weeks of the US Presidential election of 2016. The second, though less surprising is the trending climb of machine learning which is taking the form of an exponential curve. This is notable because most learning curves follow an exponential process. See the 4 laws of nature video.

Of the terms researched, the Monte Carlo method (which shares the exact same characteristics as the risk analysis search) is the only search showing relative but very long term decay in the trend. Whereas, all the other technologies listed seem to be gaining in popularity. This is in stark contrast with the results you obtain from looking at simulation as a whole (a google topic) which is very strong because it covers many domain areas that go beyond Monte Carlo.

To get a sense of what the future holds in terms of relative focus and interest in the field of analytics, we created 2 forecasts. The first of the two is our naïve forecast developed using curve fitting. The other forecast (the dotted lines) was developed using a forecasting package (Oracle Predictor) that tests, fits and ranks a multitude of time-series forecasting methods, most notably ARIMA. We ran our forecast on the relative search popularity and if the trends derived from the past are good prognosticators of interest, then machine learning is trending strongly to be the dominant focus for advanced analytics for the next 2 years.

Analytics Technologies Interest by Region/Country

If you are to take a geographical perspective, the cultural aspects seem to dictate what is popular across the globe. Big data is extremely searched and popular in both Latin America and Russia. An ironic twist given how much Russia has been in the press these days for their cyber tactics. Forecasting seems to be quite popular in western cultures such as North America, parts of Europe and Australia. There is also keen interest in forecasting in what appeared to be resource rich areas of the world including Brazil and parts of Africa which we can probably relate to resource extraction and commodity price forecasting – this of course is only a hypothesis.

Another notable observation is that big data, Monte Carlo simulation and machine learning are most popular in both China and India. Granted, this could be correlated with population size. In any case, this should be concerning to anybody in the analytical field because this means we are not alone being experts in numbers. In fact, these cultures have a long and rich history of being good with math and their curiosity should be written off only at our greatest professional peril. Instead, this curiosity and effort in the realm of science and mathematics should be more of a focus here in North America, otherwise we will be the ones left behind. Unless, of course our worst fears play out and there is a robot apocalypse then what the hell, we are all screwed anyways. But I digress.