As part of the recent big data for monitoring and evaluation workshop organised by the Development CAFÉ, our team at the Lab ran a data innovation clinic with some absolutely inspiring participants last week. Data innovation offers unprecedented opportunities for evidence-based policy-making, and has been changing the way policymakers and development practitioners address development challenges. But for many who have never been part of this process, ‘data innovation’ often comes off as ambiguous.
We approached the clinic with an understanding that data innovation means embracing new and unconventional data sources to better understand challenges and identify opportunity areas — in this case to monitor and evaluate development programmes. But in practice, how would this work?
Arming ourselves with UN Global Pulse’s Data Innovation for Development toolkit — a practical walkthrough of how development practitioners and organisations can go from idea to implementation — we spent a full day with the participants offering step-by-step guidance on how to go about designing their own data innovation project. Given the limited time, we decided to focus the clinic on three main areas: problem identification; understanding the context from which data is derived; and refining research questions. All three of these are key aspects in scoping possibilities for the use of alternative data sources for any project.
We thought the process was better captured in pictures — so here are snapshots of how things went…
Anchoring with our commuting statistics research
We kicked off the clinic with a presentation from our Full Stack Engineer, Muhammad Rheza, on how we used data from social media to reveal commuting patterns in Greater Jakarta. Rheza explained not just the results of the study but also the process of developing the research project.
Rheza’s presentation served as an anchor for us as we walked the participants through the three modules we selected from the toolkit: the Problem Definition Tool, the Data Journey Tool and the Project Concepting Tool. Throughout the clinic, we referred back to the commuting patterns study to demonstrate to participants in real-time how to use each of these modules.
First Up: Defining the Problem
Well-defined problems can lead to successful innovation projects. With the abundance of data available now, too often the focus on harvesting and analysing data that do not address the actual challenges. People often start by eyeing the solution they want, instead of first outlining the problem. Our first exercise encouraged the participants to think critically and go deeper into what problem they want to address.
Using the Problem Definition tool, we asked the participants to write down a specific development challenge that they want to address with the use of alternative data sources. They were given the choice to do so individually or in a group. They were also asked to identify contributing factors and people who are directly affected. The aim of this was to make a list of the objectives and begin looking at existing data related to the issue.
Next: Mapping the Data Journey
Naturally, it took everyone some time to define the problem — a few sketches here and there and some cases of erasing initial ideas and starting over. But once that path was clear, we moved on to the next phase — the Data Journey tool. This time, we encouraged the participants to explore the problem from a different angle: looking from the perspective of the community they are trying to help and identifying interactions that can generate data points.
For many, this meant revisiting the problem identified and looking at it with a fresh lens — that of a person affected by the problem. The Data Journey tool can also help you identify data touch points and potential research questions. The tool helps you to see whether you need to narrow the scope of the problem you want to address, or whether big data analytics can actually provide you with the insights you need to solve the problem. For some of the participants, the answer to the latter question was, “Not really.” That’s okay too — we then used the tool as a discussion starter on what sort of analysis and what types of data would provide them with the insights they needed.
And then: Project Concepting
For this final exercise, we played around with parts of the Project Concepting Tool, which can be used to synthesise ideas, refine research questions and frame hypotheses. The goal of this tool is to help participants make sure that their data innovation ideas fit the needs of the problem they’ve identified. How do we validate the results that come from analysing many of these new and alternative data sources? Similarly, these call for feasible innovative approaches. Our facilitators brainstormed with the participants, helping them to survey what current tools are out there and what could be useful for the desired outcomes.
Some pretty useful stuff here:
- People really enjoyed working with the Data Journey tool as it provides them a concrete way to understand the interaction between humans and digital platforms, and thus gain a better understanding of the context in which different digital footprints are produced.
- The commuting statistics study was really neat to have as an example — but maybe too neat, especially in terms of mapping out a data journey. Participants suggested having two case studies next time, perhaps involving a slightly more complex data journey, or a data journey where the subject is non-human. (Great idea, we’ll give it a shot!)
- Having a mix of data science and social systems people facilitating the clinic enabled us to give participants a more holistic view on how a data innovation project plays out in real life.
Many thanks to Development CAFÉ for organising this workshop and allowing us to share our data innovation toolkit, and a big shout-out to all the participants who took part! We hope to cross path again soon on our data innovation journey.
Pulse Lab Jakarta is grateful for the generous support of the Government of Australia.