Monthly Archive: March 2012

Is it too early for Open Data in Africa?

Open Data has been the hot topic over the past last few months. Open Data is an initiative that is aimed at creating a cure for the ‘data huggers’ by making data that is meant for public good open and freely available without any restrictions.

In Africa, Kenya is the pioneer of this initiative, through its Kenya Open Data Initiative (KODI) that has made over 400 government datasets publicly available through its web portal. Ghana, through the Ghana Open Data Initiative (GODI) is following closely in Kenya’s footsteps, with countries like Tunisia and Liberia looking at the same line of thought.

Now, what does Open Data mean for a country? Having been involved from about day 20 of the initial steps towards KODI, the main aim has been ways of making the government accountable and transparent while increasing citizen engagement. The government spends money in the different administrative levels eg the CDF expenditure, there are tons of statistical data collections done eg the Census of 2009 , the national budget etc etc. To some, the closest they get to this data is through the media channels of aired parliamentary proceedings.

All this data is data that should be out in the public domain, not only for researchers or analysts, but also for the public good so that the citizens are able to hold their policy makers accountable or give them a pat on the back for a job well done. The policy makers can use this data for future planning and decision making, eg when it comes to revenue and project allocation and as a ‘lesson learned’ kind of model for ways that they should not carry out future businesses but also for insights of ways that the different government organs can corporate for improved service delivery.

There are many pages of proof of concept that i can write (in my own opinion) of why we needed Open data like last decade (but i will leave that for my masters class :) ). From this, it is clear to see that we do not need to hide the data of our operation until we have perfect operations. According to article 35 in the Bill of Rights in the revised Kenyan Constitution, “every citizen has the right to access to information held by the state or any information that is held by another person and that is required for the exercise or protection of any right or fundamental freedom”. This means that, policy makers do not have to hide the ‘computer errors’, that occurred during their estimations or spending, parliament proceedings should not be just another document that is just within the parliament premises but should be a hansard (available here thanks to the Kenya Law Reporting) that is available for reference many years later.

The World Bank, which pioneered in opening up its data, was very afraid at first about opening up its operations, grants and finances data (that is available through its World Bank Finances portal) but realized the importance of opening up with the increased interaction and interrogation from the beneficiaries of their programs and an increased revenue generated from the sale of their data books due to the increased demand. This improved their impact analysis for the various projects executed.

So in my opinion, IT IS NOT TOO EARLY for Open Data in Africa. This is a necessary evil that governments should adopt to ensure citizen participation and engagement in the country development as well as transparent governments that carry out their businesses in the right way because they know, THE CITIZENS ARE WATCHING!


Data for housing agencies.

Now they say part of growing up is moving out of the nest and establishing your own.

I was onto this today. I went out looking for a house.. no, for my next home and my guide was Martin from this housing agency. we got talking to calm the process of long walks from house to house and when we got comfortable around each other, Martin told me how they have classified past house hunters and how this has influenced the way they allocate prices to new seekers. Without knowing it, Martin and his agency have been using data to their profit gains.

TV and Radio personalities/ Celebrities -Apart from the TMI that Martin has about this cluster, he says this is the worst to deal with. They are proud and often too particular. Most of these earn big salaries and hence they sell very expensively to them. He says, everytime there is a celebrity looking for a house, the house owners are starting to refuse them because most are full of drama. They take celebrities to houses they havent done businesses with before because they are always willing to take them in.

Foreigners - So Martin tells me that the reason i cant afford one of the houses that i really liked is, all my potential new neighbors are foreigners and so the houses are very highly priced. The owners expect that more foreigners can occupy the houses and they can pay because houses are really cheap for them here and in his words, “the reason why Kenyans will keep staying in shanties is because the white man can afford all the expensive and very nice places in Nairobi. Home owners are in for profit and they will raise it as high as possible when they see white skin. This is the reason why Runda, Karen, Kilimani etc are inhabitable by kenyans.”

Young girls – “So where do you work?” Martin asked me. “You know..” he continued “young ladies your age looking for the kind of house you are looking for, we mostly dont give them houses. Most of these are in relationships with politicians who are in for a good time, pay very expensively for the house for three months and then disappear and leave the girls in arrears . Home owners want to stay away from your age group”. I had to assure Martin that i will pay for the house on my own and that the closest i came to politicians is on NTV’s Bull’s Eye..

Sudanese – These ones get houses very easy. Mostly they pay very well with no complains and they also tip well. Every time Martin is approached by a Sudanese national for a house, they are always priority because they have nothing to do with their money than to live large. They are good to do business with.

West Africans and Congolese – These ones pay well just like their Sudanese friends. These, however, are into funny businesses hence they change houses very often hence they like to stay in good relations with these ones because soon, they will be looking for a new place. If they dont call in 3-4 months, Martin calls them..

Independent people – Martin put me in this class. He says those who come wanting to pay their own bills always negotiate for very good deals and they are easy to work with. However, agents dont like them because apart from the commitment fee, they never tip (yes Martin does not like me). These are good because they also always refer their “clean friends” to him.

So this is just a classic example on how different users are classified and using past data that is probably not stored anywhere and analytical capabilities of the brain, allocated resources and now for business advantage, to the agency’s competitive advantage. I would love to call this, the “theory of classification and allocation”

The challenges of Big Data in Africa.

According to Word Spy, A Data Shadow is:
n. The trackable data that a person creates by using technologies such as credit cards, cell phones, and the Internet.



Over the months, having observed the trend of data collection in Kenya and how unconnected this data is anyway. There are many challenges that face the idea of big data Africa and this post is a case study of Kenya.

Data Disconnect – Data is collected every single day in Kenya. From when you enter a bank, a building, a hospital etc. The biggest problem is there is no central connection of this data hence there is no way of verifying the data. You can enter a building and give a wrong ID number, go to a hospital and give a fake ID, go to the bank and open accounts with multiple names, this is the problem with our  data, it is not consistent! Many at times, it is very hard to profile people given their data shadows.

Data is incomplete – One of the other discussions i have had is, how do we identify persons under the age of 18? Before you get your National ID card, is there a way one can track the movement of persons under 18? There is no unique identifier for this age group and for this reason, data captured cannot be connected back to birth. There is a need for some “security number/unique identifier” etc because a person’s life is a whole life of data.

No demand for big data – Two days ago while reviewing options my insurance cover, i got a chance to talk to one of the representatives at a top insurance company in the country and what was shocking to me, with all the data that they collect, they have not been using it to identify customer trends or to improve their services. They use the old insurance model or treating each customer as a single unit.

He told me that they do not insure for pre-existing diseases and i was worried because in this country, medical records are not connected, 80% of Kenyans (or more) do not have personal doctors so how will they as an insurance company verify that the patients are not lying about their pre existing conditions? And how have they learned from past experiences?

Physical copies – One of the BIGGEST challenges in the African Big Data story is the fact that most of the data is stored in physical books and digitizing this has been a great pain. Public Hospital records over the years are in files that have to be destroyed every so often because:

  1. There is no demand for medical history for patient data
  2. Storage capacity does not allow having too many records at a go due to infrastructural limitations.
  3. High costs of digitization, that is not in the hospital’s running budgets.

This is not true for all departments though. Some departments that are very critical in research still preserve the physical files but again judging from the above 3 points, this might not last too long.

Contracts have been awarded to the ICT companies to make the digital copies, this is a good step much as  this also means that some information will be lacking but again, we all have to start somewhere right?

High costs of data storage – Apart from physical storage capacities for records, there are data centers that have been developed, noticeably by Kenya Data Networks and Safaricom, where they have data center services. It is very easy to understand that a data center needs and is not limited to:

  • Constant power supply
  • Cooling services
  • Cleaning
  • Maintenance
  • Backup (data and power)
  • etc

Now these don’t come cheap either, it has to make business sense to the provider and the user, the trouble is, most of the users don’t find a lot of value in big data, hence do not use data stores or do not digitize all together :(

How Big Data could benefit Africa

Someone once said, if you want to hide something from an African, put it in a book.

Books have been developed, turned to digital storage and the world is pretty shocked at how Africa is catching up within the global village. Maybe the problem was not with the Africans, maybe the problem was with the fact that the book was not accessible to Africans.

With the era of computers and mobile phones, information access has been made very easy to Africans far and beyond. And now there is big data. How can Africa use all the data that has been generated over the years to its advantage? My method is to reuse and recycle:

  • Big data for predictive businesses – data can be used to predict business trends and product sells. Will it or will it not survive. Customer buying and consuming trends can be used to predict whether a future business venture will or will not survive.
  • Big data for business intelligence – for a company policy to survive, it has to be supported from the highest level. For this support to take place, decisions have to be made and there are no decisions that are better than scientific decisions, those that are backed my data. In order to move forward with new company products and decisions that involve the image and profitability of a company, we are going to need more than just gut feeling or rumors.
  • Big data for targeted marketing – Customer and location profiling is something that can be done by the click of a mouse. Knowing when more females are watching tv (during Naija movies or Mexican soaps) to sell female products and when men are dominating (during football matches) to advertise male products will go a long way in sales than just advertising because that was the available slot.


  • Big data for scaling – before companies scale to other locations in this global village, information on population and spending trends of a particular region are all hidden in big data. Potential consumers can be identified from data from similar products or services in the region to be able to predict acceptable profit margins before making any commitments.
  • Big data for development -before investing in anything, governments or private companies can use big data to predict the impact of their investments. From historical events, trends, best practices, traces of corruption, sources of profits and loss are hidden. These can also be used to predict whether the beneficiaries of this investments will feel the full impact of these investments or if they would rather feel it from different angles.

All this is evident with the coming up of open data (Kenya) where the Government’s big data has been analyzed to predict areas that need extra attention in terms of poverty, eduction, health etc, politicians that might not be voted back given their CDF expenditure performance and those national parks that are most beneficiary to the county (those that are attracting the most visitors) to be able to improve those doing badly.

Again, this cannot be done by a naked mind, analytical and critical thinking capabilities must be supported from operational to institution levels.

Under Utilized Big Data in Kenya.

I hold conversations very frequently with friends about local digital content in Kenya/ Africa, and how we do not produce enough of it. Today, thinking about it, we produce just enough local digital content, we just do not store enough of it locally to have any influence, and we don’t even know it.

What this essentially means is, a lot of the content that we produce in Kenya and Africa is stored where we cannot reuse it for our development. Local digital content is basically content that is developed locally and that can be locally re-used to benefit the locals. What we are doing is locally developing international content. According to a research by Portland Communications and Tweetminster on How Africa Tweets, there are 11.5 Million geo located tweets and about 2,476,800 of Kenyans are on twitter are faithfully developing information that is stored on servers that we cannot manipulate to identify trends and visuals that can be used for local development.

Attacks on us providing content for external use aside, even those who have the capacity to use the locally developed content for business gain are not doing it. Being a step ahead especially in customer numbers has made local businesses so complacent that they are not thinking of reusing data to their competitive advantage. We have had cases of advertising to the wrong audience, making decisions to act at the wrong time and this is all because have been trained to act on gut feeling and not business intelligence.

There is a lot of information in the country, stored away in servers and computers and databases and on cloud hosts and this storage facilities have become dead end graves where we are dumping data but not exhuming it to do any further forensics on it to get further information. Profiling users in order to carry out targeted marketing, profiling business areas so as to conduct targeted product selling etc, this is all information that sits in our data stores and until we put analytical minds behind it so as to influence our decision making, we will keep making the same decisions expecting different results..