Embracing big data in Africa: approach with caution

©jezper via 123RF

From business and banking, to health, agriculture and even disaster management, the opportunities to exploit data to benefit business and societies alike appear to be infinite.

There is a dearth of research into Africa's adoption of big data. More has been written on the huge potential that big data could provide for the continent, especially around social development. In this instance, scenarios have been described about how global funders could access information at the click of a button that would otherwise take months of intensive fieldwork and research (at a considerable cost), to examples of how rural farmers could leverage important agricultural information to manage their resources.

But perhaps we are getting ahead of ourselves. While the potential of big data in Africa undoubtedly opens up opportunities, it should not be viewed as a miraculous tool that will necessarily change the face of business or the development sector. Data must be physically captured and stored before it can be analysed and constructed into meaning. The need for accurate information gathering - from the broad national level in the form of population census (new data lacking in many African countries) to ensuring sustainable technology or mobile-enabled data-collection models at the business level - is still a key link in the chain that ultimately can produce big data sets.

Data analysis

Big data refers to the manner in which data is stored and analysed, once it has been gathered. Using techniques like knowledge management and systems such as Enterprise Resource Planning (ERP) allows for powerful flexibilities in terms of analyses and the prediction of future trends. While traditional datasets consisted of rows and columns of particular information that were essentially sequential and ultimately fragmented like bits of a puzzle, now, using powerful data warehouses, a variety of different data can be stored, including multimedia objects like pictures, voice, spatial patterns, etc. These need not be sequential and the various parts can be integrated to reveal a full puzzle picture.

Integration enables parallels and links to be drawn between different types of information that might not have been comparable or had a clear association before. Not only can we analyse one's puzzle, we could analyse myriads of puzzles and find connections and similarities between them. What's more, this information can be stored remotely - in the 'cloud' - so that it is widely accessible to different stakeholders who in turn can apply multiple uses to the same pieces of information.

What this means is that business problems can be solved much faster than before, resulting in quicker and more accurate business decisions. For example, Apache Hadoop (known as Hadoop) a globally accessible, open-source big data framework that enables the handling of vast quantities of unstructured and semi-structured data, allows for speed reductions of large dataset computations from day or hours to seconds, even on handheld devices. But the technological capabilities of big data should not overshadow the fact that implementation is a business imperative. If insight and innovative thinking are not applied to a business challenge or social problem in order to make sense of, and use, vast amounts of data, the amount of data that you have access to, or how fast you can access it, becomes irrelevant .

Technological challenges

In the developed world, infrastructure costs are relatively low and this results in accessible data warehousing and data experimentation. In Africa, businesses are often faced with technological challenges, expensive connectivity and limited human resources to deal with technical issues. Arguably this could result in big data being ring-fenced as an IT issue, thus diminishing its potential to enhance the business strategy. Additionally, in the African context, there may be a steep learning curve (and associated cost), with upskilling both business stakeholders and development actors alike in understanding how to translate the exploitation of data into the long-term business imperative.

It is important to note that Africa has one of the highest Internet growth rates globally. With this goes boundless untapped and unexploited opportunities. Local stakeholders who have a keen knowledge and understanding of the local and African contexts are well-positioned to harness big data as a tool to innovate. While training and education, specifically on big data are not widely available, Unisa's Graduate School of Business Leadership (SBL) includes a focus on data warehousing (DW) in modules included in the Masters of Business Leadership (MBL) as well as the School's Short Learning Programmes.

Let us finally also acknowledge the 'now what' factor. Not even a decade ago, anecdotal information points to many local businesses and public sector departments that spent millions of Rands acquiring or building sophisticated content management systems that were often never used correctly, or at all. Data analysis is a powerful business tool. Big data reveals the rich possibilities of analysis, but is not an end in itself - the human factor is, and will continue to remain, a vital component.