It is a commonly known fact that the amount of information in the world is increasing at an almost exponential rate. Most of this information is not, however, in the form of well-structured data that computer systems of the past were explicitly designed to monitor, understand and interpret.
Most of the data created on a daily basis is what the technology savvy call unstructured data. This could be YouTube videos, word documents, web pages, audio recordings or PowerPoint presentations. While structured data tends to consist of numbers and other information that can be easily categorised and searched; unstructured data is very difficult to search and categorise.
Unless the creator of a YouTube video accurately describes the content of the video, either in the title of the video or by tagging the video with words that describe it, it would be next to impossible for anyone to find it.
In order for a computer to make sense of this kind of unstructured data it would need to stop thinking like a computer and process information more like a human being, making connections between possibly unrelated pieces of information and coming to solutions that are more human in their nature.
If you have been following the exploits of various high-powered computers over the past decade or so, you will have seen challenges where IBM’s Deep Blue beat the world’s best chess players and, earlier this year, Watson, the latest in a line of IBM supercomputers, took on human opponents at the game of Jeopardy. Jeopardy is a US game show where contestants are given the answer to a question and then they have to score points by correctly giving the question that would have led to the answer.
Shane Radford, Business Analytics and Optimisation Lead, IBM Global Business Services South Africa, explains that in order for Watson to compete against humans in a general knowledge quiz, it would not be able to access a pre-ordained set of responses as current computers do, but rather it would have to learn to create relationships between information from a vast sea of information. This would allow it to answer questions in what might be considered an intelligent way.
IBM did not, however, set out to build a computer that would be able to beat humans at their own game. The idea behind Watson is creating a computer that is able to make sense of seemingly unconnected pieces of data.
Need for problem solving
This is part of the company’s approach to analytics where the need to solve problems is tantamount. Radford explains that instead of simply looking at basic sets of data, analytical systems are built to incorporate information from a variety of sources.
This comes from an understanding that the future is not one where computer systems sit isolated and only work with data that is explicitly formatted for them. In the future, computer systems will have to be interconnected, instrumented and intelligent.
Interconnected refers to the fact that many systems will be working with each other on a number of levels, from point-of-sale systems to video surveillance systems to social network services. Instrumented refers to the number of low-level systems that are continuously feeding information into the greater information system.
These instruments may be monitoring power consumption at a manufacturing plant, the number of people entering a retail outlet or even the stock level in a vending machine.
In an isolated system the sheer volume of information that these instruments might deliver could overwhelm the ability of the system to translate that data into usable information. However, in an intelligent system all the different pieces of information are treated in context, and the system itself is able to determine the impact that one system has on all the others, creating a greater understanding of the interconnectivity of different business processes.
The final, and probably most important, element of next-generation business analytics systems is the intelligence of the system. These systems do not simply have to take a fixed set of data and output a single answer; they have to be able to provide answers to questions that haven’t even been asked yet.
‘Sources of information’
For this, the systems have to be open enough to accept new sources of information without an army of programmers descending on the system to rewrite it from scratch. He points out that the aim of this new way of processing information is vital in the age of information overload.
With all the new information it is simply impossible for humans to process everything that is created and as a result we need to create computer systems that are able to gather all the relevant information and filter it so that we can then access only the information that we need at that point in time.
By properly using this kind of technology it should be possible to tackle problems that may have seemed impossible in the past. He explains that there is a considerable amount of wasted expenditure in all organisations, but it has been very difficult to identify what expenditure has been vital and which has been excess.
By applying business analytics to the problem it will be easier to identify areas within an organisation – be it private or public — that could benefit from a closer investigation.