Building a home for government data
India has a wealth of government data. It has already published over 4 lakh data sets online, and that doesn't account for all the data still sitting on paper in ministry storage rooms.
But what about actually using this data? That's a whole other story.
Government officials, researchers, and data scientists alike agree that India's data is usually inaccessible, not user-friendly, siloed, unclear and low quality. This makes it much harder to bring vital data to bear on the country's biggest challenges.
To accelerate change and make government data accessible to both policymakers and the public, NITI Aayog announced it would build NDAP — the National Data and Analytics Platform. NDAP will be India's home for government data, where anyone can access, analyse and visualise government data sets.
NITI Aayog and their partner IDInsight reached out to us to help develop the product. One point about Design Sprints that struck a chord with them was the idea of going from product vision to high-fidelity prototypes, tested with real users, in a matter of 4 weeks. Since NDAP catered to a wide variety of users, it was critical to make sure the product experience was just right before actually building it.
The sprint team
NDAP started with an ambitious vision statement, but now it was time to turn that vision into a reality. The first step was assembling a dream team of experts on government data and policy.
We kicked off the Design Sprint with a large team of stakeholders from NITI Aayog, IDInsight, and other leading organisations. The group included team leads and associates, researchers and economists, and everyone in between.
In the past, we've always run Design Sprints in person. But thanks to COVID-19, we were all locked down at home. We moved the sprint online to Miro, which was one of our first remote sprints ever.
We also shortened our sprint sessions from full days to half days, since there's only so long you can stare at your screen.
We were afraid a remote sprint would be frustrating, but we were pleasantly surprised. We've found that people align on ideas quicker and are more efficient at prototyping and validating these ideas online. And not having to puzzle out illegible notes or digitize sprint documents was a huge plus.
Creating a home for India's government data is an ambitious task. So we started with zeroing in on what this actually meant.
NITI Aayog put out a vision document about NDAP with ideas about what it should become in 4-5 years. But what about in 4-5 months? What should the first goal for NDAP be?
As we did one-on-one discussions with the sprint team, we quickly realized that everyone imagined something different. Some people wanted to discover and view data sets, while others wanted to sort, or slice and dice, or visualize, or even do ML analysis on the data.
When we kicked off our first in-person sprint session, this was the first challenge we wanted to tackle. Our goal was to bring together all these stakeholders, highlight their different ideas for NDAP, and eventually help them agree on a path forward.
Several hours and lots of discussion later, we aligned on a set of common goals to focus on for the first version of NDAP.
After zooming in on the first version of NDAP, we zoomed back out a bit and created a 2-year goal.
We also created sprint questions, which are questions to validate about the product during user testing; and a user flow, which sets out the route that our prototype will focus on.
With the whole team in alignment, we were ready to start creating a prototype for NDAP.
We started off our second sprint session with Lightning Demos — presentations to show how other products have solved the questions we're tackling. People excitedly shared inspiration from various data.gov sites, World Bank, Wolfram, Amazon, Bing and even a hockey statistics site.
We then moved to imagining our own solutions. People who don't see themselves as artists or designer often feel hesitant to sketch, but we believe that everyone is a designer. We used a series of structured exercises to help people get their vision for NDAP out of their heads and onto paper. By the end of the day, everyone had created a complete sketch.
The next day, we put these sketches under a microscope. We spent several rounds discussing and voting on all aspects of the sketches. In the end, the team was completely in sync — every person independently voted to move forward with the same sketch!
We took that sketch and worked together to create a product storyboard. This outlined the structure and important elements for every screen in the NDAP prototype.
Prototype, test and iterate
In just one day, our designers took the sprint sketches and transformed them into a high-fidelity, interactive prototype.
We then tested the prototype with a group of researchers, which was the first audience that NDAP wanted to target.
We were excited to hear that we were on the right track. Users loved how clean the interface was, compared to other cluttered data sites. They also raved about the navigation structure and ability to merge data sets.
They also pointed out some things that weren't working. For example, we had prioritized the search bar, but it turned out the users preferred searching for data sets by sector, indicator or source. Users also asked for more complex versions of features we had intentionally simplified, such as data queries and visualizations.
That's the beauty of Design Sprints. We could figure out what users actually wanted in just a week, rather than after months of development.
Moving into development
With a detailed prototype in their hands, NITI Aayog was able to quickly create an RFP, onboard a development partner (OTSI), and start building NDAP.
At this point, we started working on two parallel tracks. On one side, development team started preparing for the product backend, data integrations and developing core workflows, Parallel started the second phase of engagement with a goal of doing a detailed design execution on the project to create high fidelity visual designs, flesh out workflows for all important use-cases and run another round of testing with both policy and research users.
From a prototype to a live design
Once development started, more challenges and questions started to emerge. Many new features had been proposed, and we needed to decide which solutions were actually important and which were leading to feature bloat. It was time to go deeper, flesh out designs for different use cases and scenarios, and further improve the visual design.
We ran a series of alignment workshops with NITI Aayog, IDInsight and OTSI to revisit NDAP's core functionalities in a lot more detail. By the end of the workshops, we identified six issues to fix during the next round of design execution:
- Making search smarter through NLP (natural language processing)
- Enhancing access to metadata
- Improving querying and filtering
- Redesigning the analysis feature to make it more powerful
- Resolving issues with the dataset-merge process
- Rethinking the post-merge data exploration and analysis
Testing it again
After several rounds of design iterations and user testing, we were able to validate NDAP's new design. Here's what users said about it:
“It’s a good starting point!”
All the bureaucrats who tested the prototype felt that NDAP would be quite helpful for them.
From vision document to live product with iterative, user-driven design
People often think of government tech as slow to build and difficult to use. We're proud to have worked with the NITI Aayog to upend this idea.
In just four weeks, we used a Design Sprint to turn an ambitious but abstract vision statement into a functional prototype that was already validated by two user groups. From rapid prototyping to continuous user testing, NDAP's development looks just like that of products from the hottest startups.
At IDInsight, we were extremely impressed by Parallel’s ability to work with a large group of government officials, researchers, and developers... We also appreciated how the Parallel team was able to immediately pivot plans to adjust to COVID-19 guidelines. Robin and his team were flexible in redesigning their systems to manage the workflow remotely and produce an outstanding product. We look forward to working more with Parallel and recommend them highly.
The Parallel team is best-in-class in India at rapidly designing complex prototypes, even when working fully remotely. They are experts at bringing together ideas and feedback from large and diverse groups of people, and making those ideas concrete... They also have strong communication skills, are always available, and know how to prioritize different asks to deliver the best product. They're fun to work with and I would recommend Parallel to anyone looking to build a design quickly!