KENYA DATA REPORT - UNDPKENYA DATA REPORT Mungai Paul University of Cape Town South Africa Abstract Kenya has received many accolades since 2011 in relation to the open data initiative.

KENYA DATA REPORT

Mungai Paul

University of Cape Town

South Africa

Abstract

Kenya has received many accolades since 2011 in relation to the open data initiative. Among this

is 2016 open data barometer report where Kenya was ranked highest in Africa. This report

describes the strategies, challenges and initiatives that define the institutionalization process and

illustrate the Kenya open data ecosystem. One of the key observations is that a national open

data initiative can start without having all the ducks in a row. However, there is need to develop

and institutionalize the right structures soon after, otherwise as was the case for Kenya, some of

the initial investment in the beginning may not yield the expected returns. This study describes

the strategies that Kenya latter developed in an effort to rekindle this initiative. Even though

these were specifically designed for the Kenyan context, they could also apply in similar

countries, or help in identifying gaps in their strategies. The report concludes with a few success

stories which help in demonstrating how open data has been adopted to create social and

economic value.

Keywords: Institutionalization, open data, transparency, accountability, law and policy, strategy.

INTRODUCTION

Kenya has been at the forefront of open data since 2011, and according to the 2016 open data

barometer report by the World Wide Web Foundation, the country was ranked at position 42 out

of 92 countries worldwide. Even though this is the highest rank among countries in Sub-Sahara

Africa, the country still faces numerous challenges in the institutionalization of open data.

Despite this shortfall, the open government data initiative still needs to be celebrated as it has

helped in creating awareness on the value of openness in the country and beyond. It is

spearheaded by the Kenya Open Data Initiative (KODI), which operates under the Kenya ICT

Authority, a state corporation under the Ministry of Information Technology. KODI has seen

enormous growth in terms of datasets and buy-in from government agencies. Currently, 902

datasets have been published from 31 out of the 62 government agencies. The open data portal

has had more than 170 million page views since its launch in 2011, and is the most visited

government portal in the country. Despite these achievements, it has had a few challenges,

especially between the year 2012 and 2013, which will be explained in the section on KODI

below. The sections below provide a description of the structures that have been put in place to

help ensure buy-in and institutionalization by the various government agencies. This includes the

challenges that the KODI has faced since it’s conception in 2011, and an account of how some of

these challenges have been tackled so far. Following this is a detailed description of the various

open data events since the initiative officially kicked-off in 2011. This includes one event that

acts as a great step towards open science in Agriculture, which entails making research data

freely available and without restrictions for use or redistribution. The final section will describe

five transformative open data projects that demonstrate the impact of official and crowdsourced

data, and the kinds of public-private partnerships that have emerged in the process. The first

project is Ma3Route, a crowdsourced solution that provides data on traffic patterns and accidents

in Kenya. The second is Google Live Traffic Alerts, which provides live traffic alerts in Kenya

through their Google Maps. The third is Newsplex, a publication by the Nation Media Group

(NMG) based on data journalism which demonstrates the use of official data for storytelling or

generation of fact-based newspaper articles. The fourth project is Outthink Urban Planning

Initiative by IBM Africa & Nairobi County, which demonstrates how data can be used to

improve urban planning in congested cities in developing states. The fifth and final initiative is

Ushahidi humanitarian model which is a Kenyan startup gone global. It relies on crowdsourced

data to give voice to marginalized communities in times of crisis, which includes natural

disasters or violence emanating from conflicts in an effort to uphold justice, transparency, and

human rights.

METHODOLOGY

This research will use the case study approach from a critical realism perspective (Bhaskar, 2008;

Easton, 2010). This approach involves definition of the research question which aims to provide

review of the open data ecosystem, address common systematic challenges that hamper the full

realization of Kenya Data Revolution and identify recommendations for follow-up actions, case

selection with a particular focus on Kenya open data initiatives, and generalizability – which will

help explain why specific complex events occur/fail to occur within the context under study.

Primary and Secondary Source of Data

Primary data will be sourced from semi-structured interviews and documents obtained from

representatives of the various government agencies, private organizations dealing with data, and

other stakeholders to these initiatives including citizens will also be identified and interviewed.

Secondary data will be sourced from media report analysis where newspaper articles, blogs, and

news feeds from television on these initiatives will be reviewed. This secondary data will help in

triangulation of findings.

Data Analysis

Data Analysis of the interviews will be through thematic analysis but guided by some of the

coding practices used in Grounded Theory research approaches i.e. a combination of open and

closed coding, coaxial coding and selective coding. Document analysis will be conducted by

means of Critical Discourse Analysis (CDA). Atlas.ti will be the qualitative software analysis

tool. The advantage of using this widely used software is that the data and analysis files can be

shared for future research with other researchers.

CASE DESCRIPTION

The Kenya Open Data Initiative (KODI) has been the cornerstone of the data revolution in

Kenya. A description of its strategies, supporting structures and challenges is provided below.

This followed by a description of the various open data events since the conception of the

initiative in 2011. This section concludes by providing a list of transformative open data

initiatives, mainly by the private sector using data from either government, crowdsourced, or a

mashup of both.

KODI Strategies and Supporting Structures

Re-launch of the KODI Portal. The KODI portal was re-launched in July 2015. This followed

enormous work on the portal to ensure continuous publication of more timely and diverse

datasets, and an improved user experience. The portal had 680 datasets at the time of the re-

launch, which was a significant rise from 262 datasets at the 2011 launch. This was as a result of

the rise in the number of active government agencies, which shifted from 4 in 2011, to 31 by the

time the launch took place. This is a change that mainly happened in 2014 and the start of 2015,

where the KODI team went on a sensitization and awareness campaign to various government

agencies to bring them on board. The rise in datasets saw a significant rise in the number of page

views, which increased by 10 million to a record high of 54 million views, making it the most

visited government portal in the country. Some of the government agencies, which were

previously publishing their data on their portals, preferred to publish at the KODI portal, since

their data got more views there compared to their website. This became an additional incentive

for agencies to submit their data to KODI.

Data fellows program. In an effort to build capacity among the various agencies and increase

data surfacing, ICTA has been contracting training and deploying sets of data fellows – mainly

young graduates to a number of government agencies to assist in build required ICT capacity in

these agencies. The contracts run for a duration of six months. Their role entails working with

the allocated agency to establish and implement internal procedures and guidelines for data

publishing. Each agency is allocated two data fellows, comprising of a computer data

management fellow and a communications/ public engagement fellow. The former is meant to

assist in manipulating and visualizing large data sets in various formats, while the latter should

assist in public communication and engagement through social media and strengthening relations

between the agency and ICTA in relation to this initiative. ICTA commissioned the first set of

data fellows in the second half of 2015, and involved 8 fellows who were embedded in three

government ministries and a county government. The second set comprises of 50 data fellows

and will target 25 government agencies. This will be a great outreach noting that there are

already 31 active government agencies.

Significant Number Blog. This is a KODI team initiative, and involves the publication of a blog

explaining a phenomenon using a significant number identified from one of the datasets. In

particular, the team identifies a dataset, analyses the data and points out some significant

numbers that they find very interesting through the blog. For example, they pointed out that the

highest number of road fatalities in Kenya occur at 6pm. This is weird because this is the time

that traffic is at a gridlock. Nairobi, the capital city of Kenya has the highest traffic and

contributes the bulk of road accidents in Kenya doing something about 50-60%. So, when we

consider that, why are all these people dying and traffic is very slow at 6 PM? (ICTA, 2016).

These are interesting questions to grapple with, and demonstrate how the value of open data can

be derived through stories. This is part of data-driven journalism, though it does not restrict non-

journalists from writing such compelling stories. The KODI team for instance does not have

trained journalists, yet they are able to come up with such compelling stories.

Request a dataset. The KODI portal has a feature that allows citizens to request for datasets.

This helps them understand the datasets that people are interested in, which they use as the

criteria to determine which government agencies to engage with in obtaining this data, and also

the agencies to prioritize for the data fellows program. Bureaucracy often slows down the

turnaround time to these requests, as the proper communication channel has to be followed. For

instance, when dealing with the ministry of education, KODI team have to write to their CEO at

ICTA, who then requests the Permanent Secretary (PS) at the ministry of ICT to write to the PS

at the ministry of education. At times once the communication gets to the other ministry, it is

directed to the wrong department, especially where they assume that it must be forwarded to the

ICT department because it’s a request from the ICT ministry, yet this may have been directed to

another office that specializes on a particular dataset. This requires the KODI team to keep

tracking these requests in order to pick up such anomalies. Their experience is that emails don’t

work between agencies, and that physical letters are the preferred mode of communication. This

is a tedious and may not be sustainable. They hope that the Access to Information bill will assist

in reducing these bureaucratic hops and that they will have a shorter turnaround time, leading to

more relevant datasets.

Data Release Form. ICTA have designed a form that helps define the contract between them and

a government agency per dataset. This provides proof that both organizations consented to the

publication and continuous update of a particular dataset. The terms and conditions are as

follows;

“1) This schedule confirms that the following listed datasets are to be published to the

open data portal, www.opendata.go.ke. 2) Acknowledges that the data will be licensed

using the creative commons 1.0 Universal framework to allow the greatest reuse potential

in the long term. 3) Confirms that the data will not be changed in any way. However

recognizes that the schema may be altered to optimize its overall usability on the Open

Data Portal. 4) It is the responsibility of the data owner to ensure they continuously

supply updated records on a regular schedule, or alternatively, 5) The Data owner agrees

to collaborate with the ICT Authority and Kenya Open Data, to phase automatic updates /

semi-automatic updates of data provided. 6) Further recommends tools or procedures as

may be constructed by both parties to support the long term sustainability of the initiative;

Including the use of Data Sync, and or other ETL tools.”

Awareness and Capacity Building Workshops. This targets government agencies, academia,

media, civil society organizations, developers and the general public to increase buy-in. These

workshops have been face-to-face, though there are plans to have webinars, as these do not cost

as much, except for the time and manpower in setting up the online meetings.

Internal Capacity building. ICTA employed more staff in 2014 and the capacity issue seems to

be addressed. The current team is complimented by the data-fellows program, which makes it

easy them to reach-out to several government agencies at the same time. They also outsource

some of the functions, for instance the human resource aspect of the data-fellows initiative.

However, this has had some drawbacks as was reported by the first set of data-fellows, who

complained of late payment and lack of resources, in particular personal computers. For the latter,

it could have been an issue of bureaucracy at ICTA, though these issues should be addressed

prior to recruitment, to guarantee productivity from the onset.

Funding. ICTA KODI project still depends on funding from the World Bank. The government

of Kenya needs to create a structure that will guarantee funding for this initiative, and also find

ways to scale the project further and accelerate its implementation. One approach is for them to

allocate budget in each government agency to ensure that each has a team of trained data

scientists, who work with the KODI team in curating and publishing data to the public. The tasks

associated with open data curation also needs to be defined within their job description and

measured through their key performance indicators, since Kenya follows the performance

contracting model. This would guarantee that effectiveness of the data-fellows program as the

right individuals within these agencies are trained.

Official vs crowdsourced data. KODI only deals with official data from government, and does

not publish other types of data, obtained by non-governmental agencies or crowdsourced from

the general public. However, the KODI team has been providing advisory and holding joint

events with other stakeholders as an effort to exchange ideas and experiences.

Expanded Partnerships. The KODI team has been making deliberate efforts to create and

maintain working relationships with the data producers who consist of government agencies, and

users who include private citizens and organizations, with the aim of ensuring continuous supply

and demand of data. They also have a partnership with Socrata, a U.S.A based company. This

followed the recommendation of the KODI taskforce and the World Bank in 2010 to deliver the

platform. This was preferred as it had the required features, and had been properly tested, and

given the short timeline between the conception of KODI and launch of the first portal, it would

have been a great stretch for a Kenyan based development house to deliver. This partnership has

been ongoing and the KODI platform got a major upgrade in July 2015, which provides an

enhanced user experience and contains more timely and diverse data.

KODI Challenges

Resistance from government agencies and the executive. Some of the government agencies still

have the notion that open data will take their jobs away or stop an already established revenue

stream, which could either be formal in the case where an institution sells its statistical

publications was the case for the Kenya National Bureau of Statistics, or informal by the staff as

tips for services rendered. In the case of the executive, it was unfortunate that open data was

likened to wikileaks in 2011 when the initiative was starting off, and there had to be convincing

by the then Permanent Secretary of ICT that the focus was on youth employment and innovation.

Government has since established the ICT Authority and set it in charge of the open data

initiative, which mainly entails supporting government agencies to publish their data to the

public in open and machine readable formats, and to provide advisory on the same by proposing

or assisting agencies to implement open data policies, and to educate the public on the value of

open data, how they can access it and how. Funding for the KODI is still based on a grant by the

World Bank, and there is hope that once this funding runs out, government will allocate budget

to support this through the ministry of ICT. This will be a great move towards institutionalization

and will guarantee sustainability. Otherwise, if the World Bank was to change its priorities today,

the initiative could be affected.

Politics: Change of the constitution, general elections and ICC cases. Kenya’s current

constitution was enacted on 27th August, 2010, replacing the old one which was in place since

independence back in 1963. It faced numerous challenges including lack of proper understanding

which resulted in misinterpretation of some clauses, which was either out of ignorance or

calculated based on the enforcers’ political interests. The constitution referendum was hotly

contested and 32% of the population voted against it, including the government of the time. This

was intensified by the upcoming general elections in 2012, where political parties used this as a

way to measure popularity and resulted in earlier and longer campaigns from 2011 all the way

until the election date, which changed from 14th August 2012 to 4th March 2013 as a result of the

new constitution. The stakes for the presidential elections were aggravated even further by the

International Criminal Court (ICC) which had indicted the current President and Deputy

President on charges of crimes against humanity. These events resulted in negative politics,

which affected government operations and saw government distancing itself from ideologies that

were linked to the west. This environment was quite unique for Kenya, having lots of internal

and external pressures and as a result, was not the most conducive for a new initiative, especially

one that is seen to touch on accountability and transparency, and being spearheaded by a

foreign/western organization, namely the World Bank. Also, the policy makers had other

interests related to elections and it was hard for them to engage in debates touching on other

important aspects such as access to information bill, which has only been ratified into law in

2016. In addition, some offices in civil service formerly held by individuals interested in politics

who had resigned as required by law, were vacant for longer than expected following the change

of election date (Obala, 2012).

Conflict of interest. Compared to countries like South Africa and Tanzania, the open data

initiative is managed by their statistical authorities. Kenya has a different approach where the

initiative rests under the ICT Authority (ICTA) and not the Kenya National Bureau of Statistics

(KNBS). These two agencies fall under different ministries, that is the ministry of ICT and the

ministry of devolution and planning respectively. This creates a challenge in communication and

collaboration following government bureaucracy. The ministries also have different goals and

priorities, and unfortunately there could be instances where there has been duplication of effort

among the two agencies. An example is the publication and/or visualization of the same datasets

on each agencies portal using different platforms, which results in different output. However,

despite this conflict of interest, over time, the two agencies have been able to curve out their

mandate and the roles are clearer now. KNBS agrees that it does not have the funding and

resources to manage the open data initiative, and instead focuses on its core mandate, which is to

collect, compile and publish official government statistics which include household-based

surveys and establishment based surveys. These statistics are not generated exclusively for

KODI, but are made available upon request. In the previous KNBS leadership, it was difficult to

obtain this data, and on some occasions before 2013, the former Permanent Secretary at the

Ministry of ICT – Prof. Bitange Ndemo had to scout at KNBS to ask for data. One particular

dataset that was obtained this way was the first batch of census data to be published on the KODI

platform. This demonstrates the need for champions at senior government management level in

requesting for policy change or sensitive data from other agencies or ministries. It is not

automatic that such a noble idea will be embraced by other agencies as there could be policies or

laws such as the official secrets act in Kenya which prevented release of some information on

national security grounds.

Complexity of data. Government departments operate in silos, follow different standards, and as

a result, present data differently and in some cases in non-machine readable formats. This has

created additional complexity to KODI as they have had to clean and convert this data in the

required formats before publishing it to the public. Through the data fellows’ initiative

(described under the KODI strategies section above), some of the agencies have been trained on

how to curate data in the required formats.

Lack of structures to support KODI. 2013 was a difficult year for the initiative. This was

aggravated by election campaigns that year and the inadequate staff capacity at KODI at the time.

This was a challenge that lasted between 2011 and 2013 where only four government agencies

were active, and the number of published datasets remained constant at 430 between 2011 and

2013. However, this changed in 2014 when ICT Authority increased the headcount at KODI,

which led to an immediate spike where an additional 70 datasets were published that year, and

the number of active government agencies grew from 4 to 25. The new entrants partially explain

the rise in number of published datasets. The rest came from datasets which had been submitted

by the already active agencies, but which did not have someone to publish. KODI has since

devised ways to sensitize government agencies to publish more data through formal structures

such as the data release form and the data fellows program which ensures skills transfer from

ICTA to the various government agencies. The passing of the accessing to information bill by

parliament and it’s assent into law by the President in August 2016 has also assisted legitimizing

the call by ICTA to other government agencies to publish their data.

Sensitive data and State Interference. Government agencies and county governments have made

an effort to release some data. However, the question on granularity still remains. An example is

county governments budget and expenditure data on the KODI portal, which are currently

generalized to sub-county’s. For instance in 2013-2014, the Kiambu county government

committed Kshs. 76 million to develop rural access roads in Ruiru sub-county. This is a massive

location in terms of land size and population, and in order for citizens to engage in depth based

on facts, there would be need to provide specifics of which roads in that location have been

earmarked for construction and the budget for each. It’s also not clear whether these are new

roads or repairs, and the type of road surface for each. Another issue is interference from

officials in government on what data to publish. For instance, the KODI portal does not have any

data on sugar production. This had previously been made available, but after the sugar scandal

erupted in 2015, there were orders from above to pull down this data. This followed

investigations of corruption and mismanagement of the sugar manufacturing company in Kenya.

There is need for policies which will make KODI more autonomous and independent, and

protected by the law for the benefit of the citizen. It is only when such data is released, that

journalists and citizens will start to find real value in open data, since conversations can be

supported with facts, and which can be presented in a court of law if need be.

Lack of awareness. The demand side of open data presents an interesting dynamic. Literacy on

how to make meaning of data, one would ask, does the ordinary citizen care about open data?

This is linked to the question on whether the state is publishing data that either affects their daily

lives or helps them keep government and their leaders in check. A case in point is the most

recent announcement in November, 2014 by the Ministry of Education to abolish student

and school ranking on exam performance. The reason given was to prevent cut-throat

competition among institutions. This new policy also affects school selection in government

institutions, favoring learners from public institutions over those from private institutions,

despite their results. This has made it difficult for parents to determine where to place their

children based on school performance. It also takes away the right of the high performers to

enjoy the hard-earned and much-deserved recognition. Parents and educators raised these

concerns through mainstream media and social media. However, almost two years later,

government is yet to revise this decision. This means a lost opportunity for everyone, including

innovators who had created mobile/web based solutions based on this data. Therefore, the issue

of awareness is not so much whether people know about open data, they are already using it,

only that they are not aware that it is called open data, and in some cases, do not know that it is

their right. We can only hope that the recently passed access to information bill will help in

overturning such policies.

Adequate OD laws and policies? The bigger challenge is whether the current laws and policies

are being implemented and not so much whether the ones we have are adequate. As it stands, the

law such as the Kenya Official Secrets Act has been interpreted in certain occasions, to favor the

decision not to release certain datasets. This becomes a difficult issue to challenge despite the

existence of laws and policies that support open data. There is need to challenge some of these

decisions in court to set a precedence, because this is the most effective way to make abstract

laws clear. This implies that there will still be occasions where certain datasets will not be

released based on this act, but the civil society must step up and challenge such decisions to

create the right precedence.

Low innovation. In 2011 and 2012, a significant number of tech innovations emerged based on

open data. However, this momentum did not to last for long due to the following reasons. First,

the model was based on incentives from the donor community – mainly the World Bank.

Developers saw this as an opportunity to make money, and the moment these funds depleted,

they moved on to the next lucrative initiative. Second, the open data portal remained dormant for

most of 2012 and 2013, with no updates or release of new datasets. This implied that applications

that had been built previously lost relevance and traction, which was a great discouragement to

the innovators. This means that despite the pullout of monetary incentives, these developers

could not monetize their solutions since official data was no longer forthcoming, and there was

no clear plan or direction, on when this would happen. It was also at a time where there was a

reshuffle in the ministry of ICT, which saw Kenya’s main champion – Prof. Bitange Ndemo

replaced as the cabinet secretary. He is the one who had promised this community of change, and

unfortunately he did not stay long enough to see his vision come to fruition. This suggests that

perhaps additional policies would have been laid during his tenure, which would have helped

maintain the momentum of this initiative, even after his departure. Recently, the most effective

innovation has been based on crowdsourced data and not official data (released by KODI), which

still faces the challenge of timeliness and reliability. Examples include Ma3Route.com which

democratizes timely transport information through crowdsourcing via social media.

Open Data Events

Kenya has held only seven open data events since the conception of the KODI in 2011. As

demonstrated in table 1 below, several stakeholders have engaged in the development of this

initiative by contributing resources and participating in the open data events. Forum on Open Data and Open Science in Agriculture, Kenya

15-18 June, 2015 Kenya Food and Agricultural Organization (FAO)

Speakers’ Organisations: Kenya ICT Centre of Excellence and Open Data (iCEOD), Network of African Science Academies (NASAC), International Council for Science – France, Kenya Education Network (KENET), Jomo Kenyatta University of Agriculture and Technology, African Journal of Food, Agriculture, Nutrition and Development (AJFAND), Kenya Agricultural & Livestock Research Organization (KALRO) Objectives:

• Provideaninterdisciplinaryforumtobrainstormonstrategiestoimprovevisibility,preservation,archivingandpermanentopenaccesstodatafromadevelopingcountryperspective.

• Createanideationsessionforsubjectmatterexperts(SMEs)tobrainstormprojectsonopendata.

Kenya National Workshop on SDGs, Roadmaps and Data Ecosystems

15-16th August, 2016

Kenya Kenya office of the Deputy President, Ministry of Devolution, Planning and Statistics and the Global Partnership for Sustainable Development

Speakers’ Organisations: Kenya office of the Deputy President, Ministry of Devolution, Planning and Statistics and the Global Partnership for Sustainable Development Objectives:

• ProvideaninterdisciplinaryforumtocreateawarenessandsensitizationonODanddiscusshowODcanassistacountryrealiseit'snationaldevelopmentgoalsandSDGs.

• Provideaninterdisciplinaryforumonknowledgegenerationandexchangeoninstitutionalandnationalinitiatives.• Demonstratethepotentialfortransformationusingopendatafordecisionmakingatnationalandsub-nationallevels.

East Africa Open Data Fest (EAODF) August, 2016 Kenya Kenya ICT Authority (ICTA), Making All Voices Count (MAVC), MapBox, Code4Africa, @iLabAfrica

Speakers’ Organisations: Kenya Office of the Deputy President, Kenya ICT Authority (ICTA), Making All Voices Count (MAVC), MapBox, Code4Kenya, @iLabAfrica, Ushahidi, University of Nairobi School of Business, City of Los Angeles, World Bank, Sierra Leone Information Commission, IBM Africa, Data Science Ltd, S.K.I.R.T.S Foundation Kenya, Tanzania Data Lab (dLab), Nigeria - Edo State Information Communication Technology Agency, GeoCensos, International Open Data Charter, Map Kibera Trust Objectives:

• Spurdemandofopendatabycreatingawarenessofopendataandacultureofopennessnationallyandregionally.• Provideaninterdisciplinaryforumtobrainstormonstrategiestoimprovevisibility,preservation,archivingandpermanentopen

accesstodatafromadevelopingcountryperspective.• Provideaninterdisciplinaryforumonknowledgegenerationandexchangeoninstitutionalandnationalinitiatives.

Open Data Civil Society Organizations’ Workshop

11 June, 2015 Kenya Kenya ICT Authority

Speakers’ Organisations: Kenya ICT Authority

Objectives: • ProvideaninterdisciplinaryforumtocreateawarenessandsensitizationonODanddiscusshowODcanassistacountryrealise

it'snationaldevelopmentgoalsandSDGs. International Workshop on Open Data for Science and Sustainability in Developing Countries

6-8 August, 2014 Kenya CODATA Task Group on Preservation of and Access to Scientific and Technical Data in/for/with Developing Countries (CODATA PASTD), Communication and Information Committee, World Federation of Engineering Organization (WFEO-CIC), Ministry of Communication and Information of Kenya(MCIK), United Nations Educational, Scientific and Cultural Organization (UNESCO)

Speakers’ Organisations: Jomo Kenyatta University of Agriculture and Technology (JKUAT), International Council for Sciences Regional Office for Africa (ICSU), World Data System of the International Council for Sciences (ICSU-WDS), The Group on Earth Observations (GEO), International Cartographic Association (ICA) Commission on Geo-information (GI) for Sustainability, International Society for Photogrammetry and Remote Sensing (ISPRS), International Society of Digital Earth (ISDE), Research Data Alliance (RDA), CODATA – China (Chinese Academy of Sciences, Beijing), USA National Committee for CODATA Objectives:

• Provideaninterdisciplinaryforumtobrainstormonstrategiestoimprovevisibility,preservation,archivingandpermanentopenaccesstodatafromadevelopingcountryperspective.

• Provideaninterdisciplinaryforumtounderstandtheneedsofacountryorcountrieswithregardtolongtermpreservation,archivingandpermanentopenaccesstodatafromadevelopingcountryperspective.

A crash course in data journalism August, 2013 Kenya Internews Kenya

Speakers’ Organisations: Internews Kenya

Objectives: • ProvideaninterdisciplinaryforumtocreateawarenessandsensitizationonODanddiscusshowODcanassistacountryrealise

it'snationaldevelopmentgoalsandSDGs. School on Machine Learning and Data Science

15 June - 17 June 2015

Nyeri, Kenya

-

Speakers’ Organisations: University of Sheffield, Dedan Kimathi University of Technology, Makerere University, UN Global Pulse

Objectives: • Showcasingandtrainingontoolsandstandardsforenhancingopendataandopenscience.

Data Science in Africa Workshop 18 June - 19 June 2015

Nyeri, Kenya

-

Speakers’ Organisations: Dedan Kimathi University of Technology, IBM Research Africa, iHub Kenya, Massachusetts Institute of Technology, University of Sheffield, Makerere University, International Centre of Insect Physiology and Ecology (ICIPE), Kenya Open Data Initiative (KODI), UN OCHA, Pulse Lab Kampala Objectives:

• ProvideaninterdisciplinaryforumtocreateawarenessandsensitizationonODanddiscusshowODcanassistacountryrealiseit'snationaldevelopmentgoalsandSDGs.

2011 Launch of Kenya Government Open Data Portal

8 July, 2011 Kenya The Ministry of Information and Communications through the Kenya ICT Board, The World Bank

Speakers’ Organisations: Kenya ICT Board, The World Bank, Open Institute

Objectives: • Spurdemandofopendatabycreatingawarenessofopendataandacultureofopennessnationallyandregionally.

• Obtainstakeholderbuy-inofopendatabydemonstratinghowitsaffordancessupporttherealizationofcorporateornationaldevelopmentgoals.

Kenya National Bureau of Statistics Data Visualization Application Launch

27th February, 2015

Kenya The World Bank

Kenya National Bureau of Statistics (KNBS), @iLabAfrica - Strathmore University Objectives:

• Spurdemandofopendatabycreatingawarenessofopendataandacultureofopennessnationallyandregionally.

Table 1: KODI Events

As observed in events such as the most recently concluded East Africa Open Data Fest and the

Kenya National Workshop on SDGs, Roadmaps and Data Ecosystems, the private sector, non-

governmental organizations and civil society organizations support this initiative. This also

includes stakeholders from other African countries and overseas. This demonstrates the

significance of these events in nurturing the open data ecosystem at national and regional level,

since they provide a great platform to initiate or strengthen collaborations, and also share

experiences and expertise.

Most of the events described in Table 1 above were aimed at creating awareness and

sensitization on OD and discuss how OD can assist a country realise it's national development

goals and SDGs. Kenya has 83 government agencies and only 31 are working with KODI team

to provide open data, the remaining 52 are yet to come on board, and this calls for further

sensitization and awareness campaigns by the KODI team. The country also has 47 county

governments, each of which have their own ministries such as health, finance and planning, ICT

and e-Government, Agriculture, livestock and fisheries. Most of this is unchartered territory

since KODI has only managed to bring on board Kiambu county. However, this may not be

marked complete as the data fellows deployed there did not report much success, following

challenges such as lack of access to data. This started with miscommunication between the

county secretary and ICT Authority, who had written a letter directed to all ministries within the

county, but which the county secretary assumed to belong to the county ICT department since it

came from the ICT Authority. This caused a lot of disconnect as most ministries were not aware

about the open data drive, and did not make necessary arrangements to work with the data

fellows. This requires a revisit of this county in the next phase of the program. Challenges

observed there could act as lessons to help improve upcoming training and support programs to

the county governments.

Related to the aim of creating awareness and sensitization on open data, ministries at national

and county level are yet to hold debates specific to their sector on how open data can help them

achieve their development goals. The only exception to this is the national Ministry of

Agriculture, Fisheries and Food Authority (AFFA), which held one workshop in June 2015 on

open data and open access with support from Food and Agriculture Organisation (FAO). This

calls for more sensitization and awareness campaigns targeting government agencies, with the

aim of spurring debate among industry stakeholders on how open data can assist them realise

their development goals.

Transformative Open Data Projects

Several transformative innovations have emerged out of open data. However, a significant

number of these innovations have been based on crowdsourced or private sector data, and not

official data, that’s being published by KODI. There is need to identify ways that official data

can be used to support technological innovations, which would use this to support valuable and

reliable information streams. This section will describe four initiatives, and demonstrate whether

these solutions have partnered with government or are using the official data.

Ma3Route

Ma3route provides data on traffic patterns and accidents in Kenya, with the aim of improving the

mobility experience of people in the major cities. It has more than 500,000 daily users who

access the platform newsfeeds either through native mobile apps, the web, twitter or Facebook.

They use crowdsourced data, which is generated by citizens while on transit. Following a report

by the National Transport and Safety Authority (NTSA), which indicated that pedestrians form

an overwhelming majority of traffic accident related deaths in Kenya, Ma3route decided to assist

in providing a solution. First, policy makers do not know the nature and location of these

accidents. Second, they rely on accident data from either the police, hospitals, insurance

companies, or the media. However, these sources do not provide all the data since not all

accidents get the attention of the mentioned parties (ma3route, 2016a).

Ma3route, in partnership with NTSA, the National Road Safety Trust (NRST), and Zusha!, a

USAID funded project by the Georgetown University Initiative on Innovation, Development and

Evaluation, started an initiative that would help generate this data, and also provide the accident

location, data which the old model does not provide. They made a call to citizens to use their

platform to help create a map of accidents in Nairobi, by using their platform to post accident

information including time, location, description and photo. There was an incentive to this where

five people would be randomly selected per week to receive a price of Ksh. 100/ $1 worth of

airtime. Ma3route then developed an algorithm to validate the correctness of this information.

They checked whether the accident had more than one report, and whether there was a photo of

the incident. At the end of six months, 7, 817 accidents were reported, 1,324 of which had been

geo-tagged and verified. The data suggested that 42.5 of these accidents happened within 500

meters of a pedestrian bridge, which indicates that they are underutilized. The accidents also

involved more public service vehicles than private cars (Agunbiade, 2016; ma3route, 2016b).

Following these findings, in an effort to obtain additional insight, ma3route made an effort to use

Kenya Red Cross and the NTSA data, which is civil society data and official data respectively.

NTSA data is available on the KODI portal, which makes it easy for Ma3route to access it

through APIs. Government agencies, also have access to Ma3Route data which they can use for

policy making and enforcement, in an effort to reduce road carnage. For instance, in the case for

pedestrians not using footbridges, government can create mechanisms to enforce their use, by

either erecting a barrier that blocks users from crossing anywhere, introducing fines, or even

deploying officers in areas known to have high accidents (Agunbiade, 2016).

Google Live Traffic Alerts

In April 2016, Google, began providing live traffic alerts in Kenya through their Google Maps’

mobile app when it is set to navigation mode. The app provides an update of whether there is

upcoming congestion, and how long you may be stuck in traffic using that route. It then suggests

alternative routes including explanations of the advantages of these options, which the user can

decide to use or follow on with an understanding of what to expect ahead. This solution is based

on a app called Waze, the world's largest community-based traffic and navigation app that it

acquired in 2013 from an Israeli based company for more than one billion dollars. It is based on

crowdsourced data from the community, from which it gathers complementary map data and

traffic information from its active users. This demonstrates that open data can transform lives,

and is also lucrative if solutions can be scaled nationally or globally, as is the case for Wazi map

which is used in many countries including the United States of America. This is a great incentive

for the developer community in Kenya, to keep generating scalable ideas and solutions based on

open data (Klosowski, 2016; Osamuyi, 2016).

Newsplex

Newsplex is a publication by the Nation Media Group (NMG) that aims at increasing depth and

creating a new image to mainstream journalism in Kenya, and address the growing need for data-

driven stories. This publication forms part of the daily nation newspaper (print and web), which

is the leading newspaper in the country. It is the pioneer in data-driven storytelling among the

mainstream media houses, and there are signs that the rest will soon follow. NMG is investing on

data journalism by supporting the Newsplex team of journalists with the required skills and

resources (Jooste, 2015; Newsplex, 2016).

Newsplex has been publishing stories that directly challenge the establishment, resulting in

policy change. Examples include a series of articles discussing why Kenya has one of the highest

incidents of road deaths in the world per capita, which challenged the National Transport and

Safety Authority (NTSA) to improve on its data collection and dissemination strategies for

public consumption. This also made NTSA revise its strategies on how to enforce traffic laws

and regulations with the aim of reducing carnage on Kenyan roads (Jooste, 2015; Newsplex,

2016).

As an illustration of how Newsplex has made use of data for storytelling, this section will

describe three articles that were randomly selected on their portal. The first article is titled

“Kenyans shackling themselves to severe debt”. It was based on 2016 national economic survey

by the Kenya National Bureau of Statistics (KNBS). This data is available to the public on both

KODI and KNBS portal. The data was analyzed in collaboration with the Institute of Economic

Affairs (IEA Kenya), a think-tank that provides a platform for informed discussions in order to

influence public policy in Kenya. Based on the same dataset. The second article is titled “The

pitfalls in Kenya’s high internet penetration”. Its findings demonstrated that Kenya’s high

Internet usage does not necessarily mean the country is the best placed in Africa to capitalize on

the Internet since far too few Kenyans can access fixed broadband internet, which is vital for

growing an innovation economy. The third article is titled “The TB patients who must sneak into

hospital”. This makes use of data from the Kenya’s ministry of health, and the World Health

Organization (WHO) 2014 Global Tuberculosis Report. It describes the trend of drug resistant

TB cases, the associated stigma, and describe how such patients should be treated (Newsplex,

2016).

Outthink Urban Planning Initiative by IBM Africa & Nairobi County

This idea was conceptualized following a discussion between the Nairobi county environment

executive and IBM Africa researchers. Mr. Evans Ondieki, the then environment executive asked

Dr. Aisha Walcott-Bryant - IBM Africa Research Scientist, how the city can best manage its

fleet of garbage collection trucks. At the time, the trucks were doing only one trip in a day, and

thus not able to collect most of the garbage in the city. Nairobi has a population of 5 million

people, and is rapidly urbanizing faster than its infrastructure development, leaving it resource

constrained. The major contributors were traffic congestion, poor road conditions, and the

inability of the supervisors to track these vehicles and devise different strategies aimed at greater

efficiency, which would then result in a cleaner city (Walcott-Bryant, 2016).

This request fitted well with the smarter cities challenge. IBM designed a scalable solution that

seeks to generate data for use in urban infrastructure planning in cities within developing

countries, and address some of the most immediate needs such as traffic congestion and safety.

The need for a new solution was premised on the argument that the current solutions, in use in

developed countries are very costly and may not work in developing countries, without

customization which adds extra costs. The current solution leverages on cheap and easily

accessible technologies, specifically mobile phones which are combined with sensors to build a

tracking device which is then installed on moving vehicles (Walcott-Bryant, 2016).

The sensor suite comprises of an accelerometer, gyroscope and a magnetometer in order to

extract not only the features that traditional fleet management companies collect, but also the

data from the device, and sending it to IBM’s big data analytics back end. The sensor suite is

able to collect data on the speed of the vehicle, altitude, road quality, location of potholes, speed

bumps and cracking along roadway networks, in addition to change in terrain from paved to dirt

roads. The sensor suites were installed in a quarter of their waste collection vehicles, and are able

to generate sufficient data to determine real time traffic conditions in the city as the trucks cover

most parts in the city (Walcott-Bryant, 2016).

IBM also developed a mobile application that was aimed at assisting garbage truck supervisors to

manage their fleet. This resulted in a more operational system that is transparent, accountable,

analyzable based on the generated data (Walcott-Bryant, 2016).

According to IBM research, Nairobi loses $600,000 in economic growth as a result of congestion

on the roads. This inefficiency is observable from the city’s garbage fleet efficiency, which prior

to the smarter city initiative, was collecting only 800 metric tonnes of garbage a day. This

number has almost doubled since the introduction of the fleet management system, and the city is

now able to collect more than 1,400 metric tonnes in a day. The city has also used the data to

identify road blockages, accidents, detours, unmarked speed bumps and hazardous potholes

enabling quicker repair time. This has assisted the city engineers and planners to design better

solutions which help in reducing the inefficiencies in the current infrastructure (Walcott-Bryant,

2016).

This data can also be part of something bigger, if it is skillfully leveraged by other businesses,

because there is a lot of entrepreneurial energy and this data might be part of something bigger.

It’s important to give people their free time, for them that’s the opportunity for them to be

themselves, to pursue the things that they want to do. So free time is everything to being an

individual (Walcott-Bryant, 2016).

This data can also be part of something bigger, if it is skillfully leveraged by other businesses or

innovators who have continuously demonstrated great entrepreneurial energy and ingenuity. The

outcome would result in more free time among city residents, which helps improve the quality of

living and support innovation, since people have the energy and time to be creative (Walcott-

Bryant, 2016).

Ushahidi Humanitarian Model

Ushahidi (Swahili word for testimony or witness) is a Kenyan non-profit startup that was formed

in response to the post-election violence that erupted soon after the December 2007 presidential

results were announced (Ushahidi, 2016a). Approximately 1,400 lives were lost and between

180,000 and 600,000 people were displaced from their homes in just 59 days (Brownsell, 2013).

In an effort to alleviate the crisis, several Kenyan software developers and bloggers joined hands

in building the first version of the platform, and according to an analysis at the time by Havard’s

Kennedy School of Government, the platform was found to be more effective than mainstream

media as it was able to report on fatal and non-fatal violence and also collate information from

rural areas. Verification of the information from eyewitnesses - which was sent through by either

email or text messages, was achieved through triangulation with additional sources from

Ushahidi partners including international media, government sources, NGOs, and Kenyan

journalists and bloggers (Ushahidi, 2016a, 2016b).

From the onset, the platform has been open source and is licensed under the Lesser General

Public License (LGPL). There has been increasing demand outside the borders of Kenya, which

began a few months after the first release. In 2008 for instance, the platform was used to track

anti-immigrant violence in South Africa, it was also used in 2009 to track pharmacy stock-outs in

several southeast African countries including Kenya, Malawi, Uganda and Zambia. During the

same period, Al Jazeera used it to collect eyewitness reports during the Gaza War between

Palestine and Israel. It has also been used to manage earthquakes in countries such as Chile and

Haiti in 2010, respond to terror attacks as was the case in Nairobi during the 2014 westgate mall

attack, track corruption related cases in partnership with organizations such as Transparency

International as was the case in Macedonia in 2011, and assist in providing realtime alerts during

other natural disasters such as winter storms as was the case in Washington D.C in 2010 or

wildfires as was the case in Russia in 2010. These are just some of the examples, Ushahidi

platform has had over 90,000 deployments worldwide, with a global outreach of more than 20

million people (Ushahidi, 2016a, 2016b).

Following this understanding, Ushahidi is best described as a crowdsourcing platform that gives

voice to marginalized communities in times of crisis such as natural disasters or violence

emanating from conflicts in an effort to uphold justice, transparency, and human rights. The

platform aggregates reports, creates visual maps, and enables direct response times, non-profits,

journalists, emergency responders, and politicians can better respond to issues happening in

times of crisis (Classy.org, 2016; Ushahidi, 2016a). Ushahidi is used to monitor elections,

generate alerts during earthquakes, and improve how people respond with humanitarian aid.

Organizations like the UN, World Vision and Aljazeera use Ushahidi to increase access to

information, protect threatened communities, and empower citizens world-wide (Ushahidi,

2016b).

They have received numerous accolades including 2016 Classy Award for Social Innovation,

2013 MacAuthor Award, 2012 Global Adaptation Index Prize, 2009 Kenya Humanitarian Open

Source Award. They have also received significant funding and support from numerous

organizations including the Omidyar Network, CISCO, Ford Foundation, Rockefeller

Foundation, MacAuthor Foundation, Knight Foundation, Google, Humanity United and USAID

(Ushahidi, 2016a).

CONCLUSION AND RECOMMENDATION

This study demonstrates that institutionalizing a country Open Data Initiative requires more than

lip service from the various stakeholders - government included. There is need for political

goodwill which should be supported by enabling laws and policies, and a dedicated resource

team whose mandate includes capacity building, awareness and sensitization. To ensure

sustainability, this initiative should be funded by government. Donor funding could help at the

initial stages, but no government should peg such an initiative on donor resources. This follows

the fact that donor organizations have their own agenda/objectives, which like any other

organization will shift with time. On the issue of laws and policies, it was observed that national

elections and change of government in Kenya affected the implementation process. While it is

expected that one government’s priorities will differ from another, such an initiative should be

shielded by law and policy. With proper structures, the life of noble government initiatives is not

sapped by change of guard or other political initiatives. An open data initiative for instance

should be in higher demand during such times, as people dig into history to create a case for

themselves, or challenge others into stepping aside for lack of results.

On the same note, the open data initiative should no longer be viewed as a tool by the west

against African governments. This initiative needs to be localized and driven more by locals than

foreigners. This implies that government and the private sector will need to join hands in

providing more support on open data initiatives including conferences and training workshops.

This calls for innovation and monetization. Government and the private sector will definitely be

involved on any initiative that has social or economic value. Perhaps the initial phases of open

data institutionalization should focus on this aspect as they seek to answer the question, where is

money in open data in country x? or better still, how can we quantify the value of an open data

initiative? Related to innovation is the need for sustainable supportive structures. In the case of

Kenya, there were numerous incentive mechanisms which saw many developers interested in

creating valuable open data-driven solutions. Unfortunately, these structures were only short-

lived, and despite the withdrawal of monetary incentives, data was not updated timeously which

significantly reduced the value and demand of the apps. Following this, developers moved on to

other initiatives, and ever since, it became an upward task to redeem their trust on the initiative.

New successful initiatives have emerged, but it would have been great to have initiatives that

began when the initiative was conceptualized in 2011.

It is therefore imperative, following the heterogeneous nature of open data, to create structures

that will guarantee timeous release of data, which follows adequate planning that factors the

complexities that may arise in curating and publishing a dataset. The approach by the Kenyan

government to have an agency set apart for the implementation of open data is a model that

should be emulated. However, it is important to ensure that other key agencies, especially the

national statistical authority are fully involved in the process. There is also need for flexibility, as

has been observed in Kenya. For instance, government agencies should be allowed to publish

their own datasets on their portal, though many as is the case in Kenya may change their

approach with time, as they realize that the national open data portal has more traffic than the

individual agency portals.

In conclusion, there is need for goodwill from government, formulation and enforcement of open

data laws and policies including a revision of laws that may contradict those supporting openness,

establishment of a government agency to manage the institutionalization of the open data

initiative - preferably a separate agency from the statistical authority, ownership by government

with less dependency on donor support, timeous release of data, and an understanding of the

social and economic value of open data which will help in driving demand and supply of open

data, and consequently guarantee the sustainability of the open data initiative.

REFERENCES

Agunbiade, T. (2016). What Ma3route Found Out About Road Accidents In Nairobi. Retrieved

November 13, 2016, from http://techcabal.com/2016/05/03/what-ma3route-found-out-

about-road-accidents-in-nairobi/

Bhaskar, R. (2008). A Realist Theory of Science. Routledge. https://doi.org/10.2307/2184170

Brownsell, J. (2013). Kenya: What went wrong in 2007? Retrieved November 13, 2016, from

http://www.aljazeera.com/indepth/features/2013/03/201333123153703492.html

Classy.org. (2016). Ushahidi Wins a 2016 Classy Award for Social Innovation. Retrieved

November 13, 2016, from https://www.youtube.com/watch?v=-CNqchiqDow

Easton, G. (2010). Critical realism in case study research. Industrial Marketing Management,

39(1), 118–128. https://doi.org/10.1016/j.indmarman.2008.06.004

ICTA. (2016). Kenya ICT Authority Blog. Retrieved November 13, 2016, from

http://www.icta.go.ke/category/blog/

Jooste, I. (2015, June 15). Data Journalism: The New Normal in Kenya. Internews. Retrieved

from https://medium.com/local-voices-global-change/data-journalism-the-new-normal-in-

kenya-213640f96408#.msc9ea7p9

Klosowski, T. (2016, February 28). Turn-by-Turn Navigation Showdown: Google Maps vs.

Waze. Retrieved November 13, 2016, from http://lifehacker.com/turn-by-turn-navigation-

showdown-google-maps-vs-waze-1761550298

ma3route. (2016a). Help Ma3Route Map Traffic Accidents! Retrieved November 13, 2016, from

https://ma3routeblog.wordpress.com/2015/04/15/help-ma3route-map-accidents/

ma3route. (2016b). Making Roads Safer with Ma3Route and DDD. Retrieved November 13,

2016, from https://ma3routeblog.wordpress.com/2016/03/15/making-roads-safer-with-

ma3route-and-ddd/

Newsplex. (2016). Newsplex. Retrieved from http://www.nation.co.ke/newsplex

Obala, L. M. (2012, May 17). The 2012-2013 Kenyan Presidential Elections. Aljazeera.

Retrieved from http://studies.aljazeera.net/en/reports/2012/05/20125171220389752.html

Osamuyi, O. (2016, April 15). Google Just Turned On Live Traffic Alerts For Maps Users In

Kenya, Nigeria And South Africa. Techcabal.com. Retrieved from

http://techcabal.com/2016/04/15/google-maps-live-traffic-alerts-nigeria-kenya-south-africa/

Ushahidi. (2016a). Ushahidi. Retrieved November 13, 2016, from ushahidi.com

Ushahidi. (2016b). Ushahidi. Retrieved November 13, 2016, from

https://en.wikipedia.org/w/index.php?title=Ushahidi&oldid=736920760

Walcott-Bryant, A. (2016). outthink urban planning. Retrieved November 13, 2016, from

https://www.ibm.com/events/ke/en/ted-outthink-urban.html