Big Data in Africa: Its Meaning

By Mbwana Alliy  |  November 23, 2014


Earlier this week I got a chance to keynote at Mammoth Business Intelligence (BI) event– the first Big Data conference in Africa over 2 days consisting of 17 speakers during the first day and hand on workshops on the 2nd day on how to build a hadoop cluster to a 101 on how to apply predictive analytical models.

Mammoth BI Interview from Saratoga on Vimeo.

Big Data and Analytics is creeping into every aspect of our lives

Since my first job out of university working in telemetry systems in UK for flight testing services support decisions around Aircraft and systems certification to later in my career as a Product Manager for Excel, Access (Office) and Business Intelligence- data and analytics interest has been a cornerstone of my career. As part of the BI team at Microsoft I saw the shift from client-server PC architecture in how enterprises were dealing with the “consumerization of IT” wave that was sweeping through the industry in mid 2000s with the rise of Google, Facebook, LinkedIn and introduction of the iPhone that has set the stage for today’s big data environment where the cost of launching a startup or setting your own big data environment has plummted allowing data enabled IT innovation to reach new industries. Today, data and analytics is both a deep discipline but innovators need to be aware of especially when combined with machine learning/Artificial Intelligence (AI) to apply in real time to create ground breaking products. We are even seeing Big Data start to disrupt how ventures are funded with the falling costs of launching startups spawning new approaches to acceleration and incubation- 500Startups’ Dave McClure would describe this as the “moneyball for startups” or “lean VC” approach.

Last month I also got a chance to attend IBM’s Think Forum which showcased the potential of IBM Watson’s intelligent computing application in being able to analyze large amounts of data and make sense of it, enabling the right questions and answers to be surfaced as needed- whether its doctors being surfaced recommendations based on large volumes of patient data and oncology journals or financial advisers being able to offer personalized investment advice in real time events unfold. An inspiring example, is how a cancer research hospital had effectively digitized patient data and oncology articles and fed this data to train IBM Watson to allow doctors to better deliver medical care [PDF]. You can see full example ideas here.


What does this all mean for Africa?
After some long discussions with some of most knowledgeable colleagues in Silicon Valley who are immersed in this new technology- I highly recommend watching this vlab to get a good overview of Deep Learning and what it means, I could not stop wondering about the applications of this technology as it becomes more accessible to REAL problems in Africa.

South Africa’s Elon Musk has recently warned about the dangers of AI. And one wonders if Africa is ready fo this kind of technology at all.  Here are some of the areas I can see big data combined and in instances, combined with machine learning can really help address important problems in Africa

  • Extending Access in Financial Services
    The context is the unbanked or formely unbanked that has currently leapfrogged traditional banking infrasture. Mobile technology has enabled financial services to reach the masses in Africa. With over $2B a month and third of Kenya’s GDP flowing through mobile money service M-PESA, Tanzania is also following fast and we can expect much of SubSaharan Africa to broadly adopt mobile money- forecasts say there will be 350M mobile money users in subsaharan Africa and the continent will lead innovation in this sector- especially when you start to think about the services that be built ontop of this data stream. Companies like MoDe and First Access are starting to do that to enable the creation of intelligent risk credit scoring that has the potential to bring sustainable lending where banks in Africa have been too cautious in past given lack of either collateral or credit history. Data becomes Africa’s newest asset. In the 2nd Afrikoin, we will explore this theme throughout the event.
  • Personalized and accessible Education.
    Africa’s education and hence competitiveness in the global economy is a major time bomb. Governments do not have the budget to increase not only access to education but quality. Private schools are filling the gap in impact investors flock to this sector as can be seen with the traction and backing that Bridge International Academies has secured- . Whilst the growth of mobile devices such as tablets will help bring content to students, there is a big data opportunity to deliver test taking and content systems that measure and adapt to students challenges and learning. I don’t its possible to be a credible Edtech company if one is not collecting and leveraging student data to help teachers personalize education, This is one of the bets we made investing in Eneza Education, test taking for elementary schools. Bridge International, whilst best known for opening many schools and education 1000s of pupils, also needs to make sense of this data as can be seen with their recent job posting for a Business Intelligence Associate in Kenya.
  • High Quality medical decisions in Rural Areas and Bringing the full potential of Genomics
    Healthcare IT is one of the most promising sectors to invest in but also one of the most challenging to find great companies. Beyond improving healthcare access and deliver to the bottom of the pyramid and middle class or improving the clinical or medical process (e.g. patient records); there is an opportunity to tap into the falling costs of genomics that generate massive amounts of data that can be used in targeting better preventitive care in areas from Parkinsons, diabetes to cancer. Fellow Tanzanian Julie Makani has made me aware of this and the opportunity to build data warhouse married with genomic research labs to gather data from African patients- check out H3Africa for more details. Whist  today services such as 22andme remain firmly in the upper levels of society, over time affordability may drop to allow African medical researchers to leverage. Big data infrastructure on the continent will be crucial in capturing, analysing this data to feel the genomic mapping gap that is emerging in Africa.


Big Data Companies in disguise- Are you building one in Africa?

Google started off as a search engine but today, many would say it is a big data and deep learning company that is able to offer personalized ads and recommendations, this is perhaps most evident when you are able to use Google Now in low cost Androids.When startups in Africa copy the Uber on demand taxi model, they often miss the fact that Uber is really a big data company in disguise of a taxi ordering app. Big data is now disrupting the taxi industry because of the way it can efficiently match and predict the demand and supply of transportation services. In areas of high demand in Africa, this makes a lot of sense, I think our investment in Africa Courier Express in Nigeria is already starting to show this promise in the logistics sector.

User Interfaces will be important in extending the power of Big Data, Analytics and Cognitive Computing to Africa

Apple’s Siri or Microsoft’s Cortana (voice recognition assistance technology) will become available to the masses as smartphone and mobile smartdevices start to become more affordable in Africa. With the massive amounts of data that can be mined via machine learning its not long before we can imagine how this might impact Africa such as allowing remote healthcare workers to be empowered to make more effective medical decisions. In the future, there will be such enourmous quantities of data that only cognitive computers and and deep machine learning techniques will be able to make sense of it all- you will not need to be data scientist or even Excel to take advantage of the data- being an ordinary African using voice recognition over a low lost connected smart devices will be enough.

Important Challenges for Africa Big Data Movement.
Throughout the theme of the different areas big data might have an effect in Africa there is the constant question one should be asking as to whether Africa is starting generate these amounts of data in the first place- one must ask whether the infrastructure and talent is in place to harness the power of big data to build relevant applications for Africa. More important than that is whether we have the right culture in place to allow this. A data analytics company founded in Kenya, backed by the same folks behind Freakonomics a year ago packed up due to a lack of demand ; some of banking and retail clients they met with did not see the value and insight the data they were collecting would provide to the bottom line. A common response was “you cannot tell me anything new about my business that I don’t already know”.
More recently, I learnt from a BBC story that the GSMA and UN had failed to convince mobile operators in the currently infected Ebola countries to release data that would allow data scientist to better analyse and map real time population movement flows that would inform high risk areas of where ebola virus might be headed- this has massive implications in coordinating resources in the right areas. This points to many organizations that are already collecting this data that might have an incentive due to both competitive and privacy reasons to have data silos or keep it proprietary. There currently is no M-PESA API but many startups depend on the data to build valuable products and services, a recent example would be the M:ledger application that was recently acquired by Safaricom. I believe M:ledger could have gone futher and added other mobile operators and banking data to deliver something closer to what 22seven is achieving in South Afrca or Mint/Intiut is delivering in the US. This points to South Africa taking an early lead in this area- from both a demand and supply side recognition of the potential of Big Data.

Tags: , , , , , ,