- In a period of Industry 4.0 automation and business intelligence, Kenya’s quest to utilise Big Data and machine learning has never been as relevant, but huge hurdles lie ahead.
- Its adoption has been slow, with a few data science solution firms in existence and only listed companies being their clients, yet it is increasingly phasing out human decision- makers from boardrooms.
- Many organisations are seeing a need for making use of Big Data with a rise in the number of data science related job advertisements increasing.
- The Government is also keen on adopting data driven decision- making in various arms of the government.
The term “Big Data” today attracts huge attention, mainly because of its power to help humans make informed decisions, turning it into an industry competition tool as new waves of business development sweep across Kenya’s corporate world.
It is helping startups, SMEs, corporates, government agencies, non-profits and individuals work more efficiently through the analysis of huge chunks of both structured and unstructured data to achieve higher operational mileage and predict future market demands.
In a period of Industry 4.0 automation and business intelligence, Kenya’s quest to utilise Big Data and machine learning has never been as relevant, but huge hurdles lie ahead.
Its adoption has been slow, with a few data science solution firms in existence and only listed companies being their clients, yet it is increasingly phasing out human decision- makers from boardrooms.
“We are seeing a huge market demand in the use of data tools by several companies and government parastatals which cannot be satisfied by the current rate of supply,” notes Mr Timothy Oriedo, founder of Big Data training firm, Predictive Analytics Lab.
“There is low understanding of Big Data and acceptance into business strategy. Converting it into insights has been a challenge for many organisations. May companies are clueless on the kind of value they will get in these insights,” says Mr Andrew Mukabana, a data scientist.
He adds that adoption of data science requires quite a substantial investment. “While many organisations want to be on this path, budget constraints are becoming a hindrance especially among startups and SMES.”
This has made the gap in the provision of data science training services wider, as the amount of data to be analysed keeps surging. But experts are optimistic of the future.
Computer Science lecturer at Maseno University’s School of Computing and Informatics Lilian Wanzare says upcoming learning institutions offering courses in data science are breathing hope into the field.
“Predictive Analytics, Moringa School, KCA university and Strathmore University’s iLabAfrica offer certification in data science. Several meetups also exist where people learn and share knowledge about the technology,” she says.
Associate professor at the University of Nairobi’s Business School Bitange Ndemo says online portals could offer huge training opportunities for Kenyans seeking technological innovations.
“Online training makes it more accessible for anyone because most of the trainers offer it for free. Every professional worth of his salt needs these skills, and online is the best way, especially when you have a busy schedule that cannot allow you in class,” he says.
But many employers will require you to produce a data science certification to prove that you have undergone training and understood the field.
Dr Wanzare identifies online courses such as Coursera, Edx, Udacity and IBM as the global popular ways of gaining knowledge and certification.
“They are internationally accredited. At a small fee, one can join a community of online learners and benefit from experienced mentors,” she says.
But Barclays Bank (Absa) Chief Data Officer Hartnell Ndungi warns companies against employing data scientist without vetting them thoroughly.
“HR departments must stop running on the Big Data certification hype to hire people. Nearly 60 percent of data scientists who turned up for my interviews in 2019 relied on the hype of Big Data. They were unable to deliver the job obligations.
“Extremely qualified data scientists are expensive. As you hire, it is wise you give candidates real life scenarios and problems to solve to avoid employing people who get raw information from the web and lie on their LinkedIn profiles,” he advises.
He asked corporates to scale up employees through internal training, using either decentralised, centralised or federated modes of on-the-job training.
“HR professionals must be the first employees to take up data science courses so they are able to lead the training policies. They must identify which employees qualify to be moved to data departments for training,” says the expert who has been in the industry for 11 years.
Mr Oriedo emphasises the need for Kenyans to get training from providers in the local market as online trainers use data sets that don’t apply to the Kenyan business atmosphere.
“We don’t need foreigners to teach us how to analyse data from our own economy. Kenya needs native data science trainers who understand local data sets to make its comprehension easier.
“Foreign experts have little knowledge on the socio-political and economic issues happening locally. When you train, the aim is always to make everyone understand, and using local examples is the best way to achieve this,” he says.
Many organisations are seeing a need for making use of Big Data with a rise in the number of data science related job advertisements increasing. The Government is also keen on adopting data driven decision- making in various arms of the government.
So, how does the future look like?
“With this trend, we expect to see new opportunities emerging because of the insights from data. In 2020, I look at SMEs as one of the main beneficiaries of Big Data in Kenya in terms of creating new growth opportunities for them as listed companies become more data- driven. An organisation’s ability to properly utilise Big Data will break or make it,”says Dr Wanzare.
Government entities and telcos keep on receiving data from various sources, but this data is not integrated. This is challenging when it comes to implementing Big Data solutions in the security, health, transport, education, finance, agriculture, communication, tourism and even sports. In his technical paper published by the Center for Global Development (CGD), "A Farewell to Disruption in a Post-Platform World", the founder and president of mPedigree,- Bright Simons, says that data hyper-integration is replacing disruption as the dominant motif, riding on the back of the triumph of integrations over data and algorithms.
“The motivation for hyper-integration has come from obvious glaring gaps in the tech world. Risk arising from compounding externalities; fraud emanating from the rapid speed, scope, and scale of transactions; performance resulting from growing complexity of requirements; and coverage due to the desperate search for top-line revenue growth and market ubiquity,” he said.
Corporates use standard enterprise resource planning (ERP) software to keep stock of core business transactions such as accounting and human resource. All these systems store data in their respective databases.
However, this presents challenges about how a firm can reconcile transactions. There are so many links to be made between the data in various systems. Worse still, the systems are from different vendors, meaning they use different descriptions for similar things. If ever you are going to see the big picture, all the data from all these separate systems need to be harmonised.
A big hindrance in the Big Data industry is harmonisation of systems and rapid digital transformation. Done well, this shall eventually help firms become adaptive and plug in systems that capture data seamlessly.
But even with a new data law being enacted in Kenya, an expectation for companies to respect customer data is still low, as the nation grapplesing with the challenge of a lack of a national data centre.
It is a deterrence to data democracy, as was witnessed during the 2017 presidential election results, where a lack of a data centre tied the hands of the electoral commission in opening the servers which were hosted outside the continent. You only have authority to open servers if they are hosted in your country.
“The main reason for having a data centre is to own a centralised location for hosting it. It is centrally collected, stored and processed. Organisations can have a dedicated data centre, a shared data centre across organisations or cloud-based ones,” says Dr Wanzare.
A national data centre would provide a centralised facility for organisations and government entities.
“So far only big companies can maintain their own data centres because of the operational costs of having a dedicated one. Small organisations and startups have to rely on cloud-based services and sometimes one is not very sure where their data is being hosted. Issues arise when there is a privacy breach and the data is located in another country,” she says.
The new data law complies with the European Union’s General Data Protection Regulation. It was a critical legislative landmark, not only in better handling of citizen data, but also in encouraging businesses and investments from international organisations that have comply with similar stringent data laws.
“Data storage and integration demand is high now but firms will have to embrace integrity. You have to be more careful now, as there is a legal framework to punish those who misuse user data,” says Prof Ndemo but warns against the cyber insecurity that comes with it.
It is an organisation’s mandate to properly secure the data they collect especially if it contains personal identifiable information.
“For organisations, the tips revolve around the technology used, policies in place for data security, training personnel on security issues, monitoring data movement for possible breaches,” notes Dr Wanzare.
Cyber intelligence has been identified as a sure way of being a step ahead of cyber attackers by scouting for malicious leads and analysing them to thwart cyberattacks in real time, according to Mr Niall MacLeod, Director of Solutions Architecture in Europe, Middle East and Africa at global threat intelligence leader Anomali.
“Just like police intelligence, cyber threat intelligence monitors potential sources of security breach, evaluates, analyses them and advises IT security departments on best actions. This process requires more data science skills as Machine Learning is the key driver for accurate analysis that reduces false positives,” he told In-Depth.
For big data hosted on clouds, organisations need pre-emptive measures to ensure a 99.99 percent of uptime, and recover fast enough once an attack lands on their networks.
But within these complexities lie opportunities to foster economic growth. There is a totally new thinking as business disruption is now anchored on big data analytics. “Old problems are being solved using new techniques. This has brought about economic improvement by lowering the cost of consultation to inform decisions. And the need to stay ahead of competitors is now relying on big data. But the overarching message for 2020 will be to understand the ecosystem first,” says Mr Mukabana.