Given the hype, let me start by explaining the buzzwords in the title of this article. What is Big Data? What is Analytics? How about Machine Learning? What is Artificial Intelligence?
More importantly, why should Ghana and Africa care about these ideas now? These are all legitimate questions you should ask. I’ll attempt to summarize the meaning of these words without bugging you with too many technical details.
In simple terms, Analytics is the process of transforming data into insights to make better decisions. Analytics can be descriptive, diagnostic, predictive, and prescriptive. Descriptive analytics answers the question of ‘what’ has happened while diagnostic analytics goes further to explain the ‘why’ behind what has happened. Predictive analytics goes one more step to combine the insights from what has happened (descriptive) and why it happened (diagnostics) to answer what will happen in the future.
Finally, prescriptive analytics aims to put all the above together and makes a statement of what we should do to get the optimal business outcome. All these processes or tasks take place within the framework of the business, and we are not talking about any theory here. Take the business of loan management for example. Descriptive analytics here could answer questions such as how many customers were granted loans in the past six months and what percentage of these customers have defaulted on their monthly loan repayment. Diagnostics analytics will go further to understand the root cause of why some customers are not paying their loans (i.e., the determinants of loan default). These root causes may be the loss of jobs, family situations or high interest on the loans.
Given these insights from descriptive and diagnostics analytics, the loan company can employ predictive analytics to predict the likelihood of customers defaulting on their loan payment based on the historical data they’ve collected on existing customers. Once these predictions can be made for all incoming customers, prescriptive analytics will then be employed to grant loans to customers that meet a specific profile based on the predicted default rate of those customers (and perhaps other business factors). This is just a high-level example to illustrate my main point about how analytics can enhance decision making.
Analytics has been in existence for many decades. In most developed countries, the mid-1950s marked the beginning of what most practitioners refer to as the “real” beginning of analytics for businesses. Before the 1950s, analytics was predominantly used for academic, government and scientific research. With the focus on basic reporting and descriptive analytics during those days, a very small group of people were typically isolated to the backrooms where they crunch numbers to report performances to support basic internal decision making.
Fast forward a few more decades and people would start hearing more about analytics and analytical software companies such as SAS came up with advanced analytical techniques and packages that can be executed on relatively larger datasets. Even during those times, very few organizations and individuals had the skills to make sense of historical data to understand what has happened in the past.
Only big IT organizations could query structured databases to pull historical data and report findings. In those days, not all data may be stored due to the high cost of storage capacities. The computational limitations of the computers and software at the time also impeded the analytics in several ways, for instance, by restricting the analysis time frame to a few months.
However, the story has changed completely in recent times. Analytics has become so popular in many parts of the world, with analytics professionals becoming one of the top paid employees in the US and the UK. The October 2018 edition of The Economist notes that, “A branch of AI known as deep learning, which uses neural networks to churn through large volumes of data looking for patterns, has proven so useful that skilled practitioners can command high six-figure salaries to build software for Amazon, Apple, Facebook and Google.
The top names can earn over $1m a year” (see the article titled “New schemes teach the masses to learn AI”). Several businesses now have in-house analytics team with several analytics professionals, and many of them having C-level analytics leaders. This is largely due to the advancement and integration of information technology into business and government processes which tracks and produces lots of data – Big Data. Businesses now need to make sense of those data to bring insights and value to their organizations.
The terms Big Data, Machine Learning and Artificial Intelligence all became popular following the advancement in computer technologies producing new types of data that required new techniques to analyze them. Big data refers to large volumes of structured and unstructured data that are produced each day with activities of businesses and individuals. Structured data refers to traditional data that have pre-defined formats and resides in relational database systems and spreadsheets such as transactions made by customers, library catalogues, economic indicators and the like.
Non-structured data, on the other hand, refers to data that do not have pre-defined formats such as email messages, word processing documents, videos, photos, audio files, web pages and other kinds of business documents. With all these big data at hand, there is the need to analyze them and traditional analytical techniques can’t handle most of these types of data.
Artificial Intelligence and Machine Learning provide the breakthrough! Artificial intelligence is a broader concept that refers to the ability of a machine or a computer to think, learn and therefore act in smart ways. Machine learning is an application of artificial intelligence that provides computers the ability to automatically analyze, learn and improve from experience without being explicitly programmed. These are the new ways by which analytics can be conducted on big data to yield interesting insights.
The good news is that the current computing architecture does allow for all these analytics to be done very quickly. The days where analytical software can only be implemented on expensive physical servers by one expert who produces reports for the entire organization are over.
In today’s world, even mobile phone operating systems can run analytics software and conduct complicated computations within seconds. Access to analytics software has become so easy and with open source software and cloud computing, some of these analytics and machine learning software come for free. If you are a business in the western world today with no analytics behind your strategies and operations, you could definitely go out of business within seconds.
In many countries, analytics has played significant roles and redefined operations for higher efficiencies and profitability in many industries, including but not limited to customer relationship management, banking and finance, insurance, fraud detection, social networking, human resources, energy management, supply chain management, pharmaceuticals, health and many government institutions.
In my many years of analytics practice, I have seen analytics transform unprofitable businesses to very profitable entities So here’s my question: can Ghana also take full advantage of the power of technology and analytics to transform businesses and develop the economy? My answer is a conditional yes. Considering the recent emerging technologies in Ghana and their integration into business processes, Ghana and Africa can benefit from using analytics and machine learning to bring insights to businesses if the following measures are implemented.
First, we need to develop the right skills to handle an end-to-end data management strategy in a complex business environment. This includes training on modern ways of storing data in a way that analysts can easily query and analyze. Understanding Big Data storage and heavy computations in a cloud environment are other ways that most successful companies are approaching their data management challenges.
Second, we should develop the right talents that will not only create reports from the data but completely transform them into actionable insights by leveraging all types of analytics (descriptive, diagnostics, predictive and prescriptive) and using machine learning methods where appropriate.
Higher educational institutions have a big role to play in this area. The structure of courses offered in the departments of Computer Science, Statistics, Engineering, Business and other related disciplines need to be reviewed to equip students with skills that can help graduates from these disciplines handle the massive amount of data to provide insights and add value to businesses.
Finally, business leaders should embrace analytics by identifying areas within the business that analytics can play critical roles and invest in the right human resources, tools and infrastructure to completely implement end-to-end analytics and machine learning solutions.
Written by: Augustine Denteh, PhD and Delali Agbenyegah
Augustine is a Postdoctoral Research Fellow in Econometrics in the Department of Health Care Policy at Harvard Medical School. He can be reached at email@example.com
Delali Agbenyegah is the head of Data Science and Analytics team at Express, a large US based Fashion Retailer where he leads a team of Data Scientists In developing and deploying machine learning and optimization models for Express.
The views expressed in this article do not reflect those of Express or any of Delali’s affiliates. Delali can be reached at firstname.lastname@example.org