Case study: Tanzanian Data Lab (dLab)

The ODI supports programmes – in the UK and internationally – that build skills and expertise in data, in support of its vision for people, organisations and communities to use data to make better decisions and be protected from any harmful impacts.

One of the ODI’s manifesto points is around capability: ‘everyone must have the opportunity to understand how data can be and is being used. We need data literacy for all, data science skills, and experience using data to help solve problems.’

The Tanzania Data Lab (dLab) programme was set up to improve data skills in Tanzania, and support the national interest in using data to make better decisions and progress towards the United Nations Sustainable Development Goals.

The dLab is an open physical space, for members of the data community in Tanzania to meet, participate in training, and engage in community events. Its aims are to help to enable and foster innovative uses of open data, connect data with policymakers and groups working towards national priorities and global goals – from civil society organisations to academia to businesses – and, ultimately, strengthen the data ecosystem.

The dLab was launched with the help of the ODI, working in partnership with the University of Dar es Salaam, SB Consulting (SBC4D), and Intrahealth, with support from the Millennium Challenge Corporation (MCC) and The United States President’s Emergency Plan for AIDS Relief (PEPFAR).

Through this initiative ODI has supported the technical and organisational capacity in Tanzania to develop the data skills and experience needed in the country.

The dLab is part of a broader initiative, Data Collaboratives for Local Impact (DCLI), which aims to strengthen the availability and use of data to improve lives, and empower citizens to hold governments and donors more accountable for results – especially in relation to HIV/AIDS, global health, and gender equality.

Facts and figures

Between April 2016 and June 2018:

• 10 courses were developed by dLab with the support of ODI trainers, in topics including data visualisation and analysis, data cleaning, data management, and publishing open data
• 1,043 trainees were trained at dLab, over half of which were women and over 300 were trained in data management and data innovation through the ‘Women in Data’ programme, which the ODI helped to design
• 18 regions across Tanzania were reached by the programme: Arusha, Dar es Salaam, Dodoma, Kilimanjaro, Lindi, Manyara, Mbeya, Morogoro, Mwanza, Shinyanga, Singida, Tabora, Tanga, and Zanzibar (across its five regions)
• 70% of all trainees reported acquiring new knowledge from workshops, while 76% of all women reported the same
• 908 people visited the data portal
• 288 datasets were downloaded
• 268 organisations were reached through community events, such as policy roundtables, workshops, and the Data Tamasha data festival
• 22 workshops were delivered by ODI-trained dLab trainers to 598 trainees (2 Data Management, 3 Data Visualisation, 4 Data Innovation, 7 Open Data with Data Management, 1 Data Publishing, 1 Storytelling, 4 Open Data)
• 3 events (bootcamps/hackathons/mapathons) were created
• 11 policy roundtables were organised and run by the dLab to discuss data gaps, progress and opportunities of data use for achieving national development priorities

Why is data literacy in Tanzania important?

With ever-increasing amounts of data being created around the world, governments, NGOs and civil society groups need skills to leverage data in policymaking, decision making and economic development.

However, data literacy in sub-Saharan Africa has generally been low without much in the way of spaces or support for data training and innovation. This has undermined the potential that data-enabled innovation could have had for social and economic benefit in the region.

While mounting international effort is put towards monitoring and measuring progress made towards the SDGs, there is great potential for data to also be used to target interventions that address the problems. Improved data literacy can encourage better use of data and decision making in policy innovation, business and academia, which together can work towards improving development outcomes and quality of life more broadly.

Data-related commitments

The SDGs were launched by the United Nations in 2015 in order to improve human life across the world in domains such as education, healthcare, and the environment – particularly in low-income countries.

The DCLI programme was set up by MCC and PEPFAR to help make progress towards and measure the SDGs, by increasing country-level use of data to improve programmes and policies; and to improve the capacity of individuals, communities and organisations to use data to solve problems relating to HIV/AIDS, global health, gender equality, and economic growth.

The vision was to:

• Enable and showcase the effective uses of data about HIV/AIDS, global health, gender equality and economic growth by government, citizens and other stakeholders to enhance decisions and innovation
• Improve policy making and service delivery
• Ensure citizen feedback loops are augmenting transparency and mutual accountability for results.

The ODI approach

The ODI was brought in to support the implementation of the dLab with training – both through the training of trainers (ToT), and some follow-on training in collaboration with those by-then ODI-certified trainers. The ODI also supported the newly-created dLab in the development of a business plan and a monitoring, evaluation and learning (MEL) system.

Aims

The ODI sought to bring its expertise in strengthening data ecosystems through developing data literacy programmes as part of the dLab project. The ODI believes that building data literacy for all is essential to build capability to use data intelligently, and to empower citizens.

Over the past six years, the ODI has developed a range of learning products and curriculum for the dLab project, focused on helping individuals, organisations and communities to get value from data. This includes a ToT programme (part-funded by the World Bank), face-to-face workshops, and online learning.

By partnering with local actors such as the University of Dar es Salaam, the ODI sought to contextualise these methodologies and materials to the Tanzanian context, while building the capabilities of a core team of trainers and data scientists to continue to deliver activities in the long-term. The ODI finds this approach to be more sustainable and suitable to local audiences.

The importance of data infrastructure

Robust data infrastructure – data assets, the organisations that operate and maintain them, and standards, guidance and policies describing how to use and manage the data – provides the foundation for effective data-enabled tools and services. These can support evidence-informed decisions which in turn generates impact from data.

What did we do?

The ODI partnered with SBC4D, the University of Dar es Salaam, and Intrahealth. Key activities included training a group of data champions who could go on to run training throughout the country, training for government staff, civil society groups, and businesses on how to open, publish and use data for decision-making.

As the ‘learning partner’, the ODI facilitated ToT, delivered trainings for end users, and continued to support local data champions through mentoring and co-design of learning tools and curricula (including eLearning).

The ODI also advised on the business development plan a ‘women in data’ campaign to expand activities to more women and girls; and a monitoring and evaluation (MEL) system (including indicators for outputs such as numbers of persons trained, and post-training feedback and skills retention), to promote sustainability and to track progress and impact.

What was the impact?

Through its training programme, the dLab has helped to improve data literacy – and the potential for better use of data in decision-making – of data users from 18 regions across Tanzania. Of over 1,000 trainees, over half were women and over 300 of those were trained in data management and data innovation through the Women in Data Programme.

The dLab helped build the data economy of Tanzania by hosting workshops, improving the national data portal, and organising events like bootcamps, roundtables, and hackathons. Many important nationwide services have used the dLab to improve their data skills to get better results.

Examples of these services include Ramani Huria, the effort from the local OpenStreetMap team to create accurate maps of flood-prone areas, and visualisations of C-Sema Tanzanian Child Helpline data from Save the Children to help combat neglect, abuse, and violence against children.

In conjunction with partner organisation, Data Zetu, the dLab also delivered a data management course to The Mango Tree, to improve its use of data in its mission of supporting orphaned and vulnerable children in healthcare and education matters.

Some of the most successful data users trained at dLab were young women and girls. Over the course of 2017–2018 the dLab hosted social enterprises such as She Codes for Change and Apps and Girls, which focused on delivering training to aspiring women social entrepreneurs. One dLab trainee, Kokubanza Timanywa, 18 at the time, created a platform called Vitabushelf for textbook exchange after noticing that students were struggling in school as they could not afford textbooks. Asha Abbas, another trainee who was also 18, used the skills she learned at dLab to improve her project Aurateen, which helped provide a sexual health forum for teenagers in Tanzania.

The dLab project was also independently reviewed with the reviewing stating the impact of the project was:

Lessons and challenges

During the course of the dLab project, we learnt that ODI can very successfully support the ongoing development of capabilities in data in an international setting through appropriate support to local partners.

Several aspects and approaches in particular worked well in our training programme.

1. Being selective about trainers to train
   We found that an effective way to develop capacity of local delivery partners and intermediaries was through a highly selective ToT approach, followed by sustained coaching/mentoring, and co-delivering work while progressively transferring responsibilities to local partners. Doing some preliminary training and ensuring that trainees had some technical background made it more likely that they would retain the skills they learnt and go on to help organisations improve their data practices and decision making.
2. Making learning tools accessible and relevant to local trainees
   Following the training in Tanzania, all trainees who were interviewed said they appreciated the teaching methodologies they were introduced to during the ODI’s ToT programme. The tools – such as OpenRefine and Tableau – and datasets being used to teach were particularly accessible and relevant to the trainers, which increased the likelihood of uptake.
   
3. Customising or localising training materials and tools
   We found that it was necessary to adapt the ODI’s general training approach to meet local needs and learning preferences. We achieved this through additional practice sessions, one-on-one contact points, and virtual mentoring/technical backstopping. This sustained level of engagement helps to ensure a legacy of local capacity that will continue beyond an individual project.

There were various challenges throughout the programme which helped the ODI to learn from what worked less well.

1. A shifting political environment can affect the national appetite for open data.
   As the dLab programme was implemented, there was a change in Tanzanian government policy and perspectives towards open data, meaning that government officials who might have been enthusiastic about applying new open data skills to their jobs tended not to be able to do so. The change in political support for open data meant the programme as a whole had to pivot the focus of trainings to the demand for data and other civil society and private sector data stewards. The desired impacts on policymaking and budget/resource planning based on data for the SDGs were not achieved.
   
2. Business planning for sustainability takes time.
   While reasonable business and MEL plans were put in place in Tanzania with the ODI’s assistance, more effort and resource would have been needed to ensure they were widely understood and acted upon after the formal partnership finished.
   
3. Indicators must be selected and monitored carefully.
   The implemented MEL system reflected the needs of the funder and had plenty of indicators but did not sufficiently cover the organisational level impact needed for sustainability. Impact needs to be measured both at the outcome and organisational levels to enable impact assessments to inform planning and to provide evidence to new funder. Indicators should be selected carefully and regular learning feedback loops need to be planned for and facilitated.
   
4. Introducing online platforms and tools needs careful support.
   The online aspect of the dLab training programme was less effective than the face-to-face training component. Use of the online modules was patchy, and there were some challenges in handing over ongoing responsibility for the system. It is unclear whether this is a specific challenge in that particular context, or a programme-wide issue, but it would have been beneficial to focus more on equipping partner organisations to enable them to administer the online modules.

Recommendations

For others implementing similar projects involving international training partnerships and skills transfer, we would recommend the following.

• Select appropriate local partner organisations with the capability (or potential capability) to conduct the bulk of on-the-ground implementation and convening. In the dLab case, the University of Dar es Salaam partners proved a terrific addition given their underlying technical knowledge and familiarity with teaching, which provided a strong starting point.
  
• Vett for qualified trainees in ToT programmes with the aim to select those who have some IT literacy and knowledge of ‘open’ principles prior to participating. It is also very useful if they have sufficient seniority to apply new skills in their work, such as at senior management level.
  
• Fully equip partner organisations to run and work with online modules, to minimise dependence over the longer term.
  

Ensure sufficient time is invested into advising on business plans and MEL-system development, ensuring appropriate impact indicators are collected to aid accurate evaluation for fundraising and reporting purposes.

Conclusions

The ODI’s support to the dLab initiative has proved to be very successful in developing the ability of partner organisations in Tanzania, who continue to develop the data skills and experience of others within the country, to help solve problems.