At CivicDataLab, we enable changemakers to leverage data and responsible AI to drive public good. Bringing together data, technology, design and social science, our work begins with a simple but fundamental question: how can better and responsible use of data and AI help public resources reach the people and places that need them most?
With our work in Assam, India, that question led us to examine the relationship between maternal and child health outcomes and public procurement. In 2022, our analysis of 165 procurement tenders revealed a striking pattern: some districts with the highest maternal and infant mortality rates also recorded some of the lowest procurement activity for maternal and child health infrastructure. The analysis exposed a broader challenge – not an absence of data but the lack of systems that connect it, make it visible and translate it into decisions that reach vulnerable communities.
This challenge is particularly acute in Assam, one of India’s most climate-vulnerable states, where health, climate and socio-economic risks increasingly intersect. Having worked with the Government of Assam since 2018, we began asking a broader question: what if, instead of building another standalone health solution, we built the digital data infrastructure needed to connect women’s health outcomes with the government budgets, public procurement data, climate risks and socio-economic conditions that shape them?
Data exists, but doesn’t connect
Women’s health data is not scarce. It exists across hospitals, government programmes, surveys, digital platforms and research institutions. The challenge is that these datasets sit in silos, so the questions that matter go unanswered. Where are maternal health outcomes improving, and why? Which districts need greater investment in reproductive care? How do climate risk, nutrition and healthcare access shape women's wellbeing together?
Our solution, FemHealth Data Collaborative, builds the connective infrastructure that brings these pieces together, so governments can understand the broad conditions shaping women’s and girls’ health. Assam is our starting point, but the ambition is to demonstrate how open, interoperable data infrastructure can support more equitable and evidence-driven public decision-making.
FemHealth Data Collaborative maps, measures and tracks critical datasets and indicators bringing together governments, researchers, development partners and civil society organisations. It has four interconnected components:
- An open-source FemHealth Data Platform consolidating datasets across health, climate, public finance, procurement and socioeconomic domains, built on CivicDataSpace, our open digital public infrastructure.
- AI and machine learning-driven analytics. Natural language processing, large language models (LLMs) and spatial analysis extract indicators from budget documents and tender notices that are prose before they are data; spatial clustering models hotspots where low service uptake and high vulnerability coincide; and prompt-based LLMs turn outputs into summaries in three languages.
- Capacity building and institutional reforms, training officials in data-driven decision making and embedding a gender lens into public finance management.
Peer learning and collaborations that develop interoperable datasets and strengthen the ecosystem together.
For us, open source is not a licensing decision but a governance position. We are asking departments to share data about vulnerable populations, and they are entitled to understand what happens to it. An open approach allows institutions to understand how data is transformed and used, enables researchers and civil society to examine and improve the methods, and gives governments greater ownership over the systems they adopt. It also makes it easier for what we develop in Assam to be adapted and reused elsewhere.
Where we are now
Our work with the Government of Assam provides an important institutional foundation for FemHealth Data Collaborative, meaning we are building within existing government systems rather than alongside them.
As part of the inaugural UNICEF Femtech Ventures cohort, we are taking the FemHealth Data Collaborative forward in Assam by co-creating an open-source platform, developing analytics to help decision-makers identify needs and resource gaps, and strengthening institutional capacity for gender-responsive use of data in public planning and finance. A beta platform is already under development. Assam is the pilot but the architecture is designed to be modular, open source and adaptable.
Importantly, this is not simply an investment in a technology product. It is an investment in the ecosystem around it: the people, partnerships, standards and institutional processes required for data to translate into better decisions.
Women's health cannot be transformed by technology alone. It requires institutions that can see the full picture and act on it.
By institutionalising gender-responsive data use within Assam's public finance management systems, FemHealth Data Collaborative aims to help governments allocate resources more effectively, identify emerging risks earlier and strengthen coordination across departments.
We hope to demonstrate that improving women’s and girls’ health begins long before she enters a clinic. It begins with ensuring that her realities are visible in the data, reflected in public budgets and considered in the decisions governments make every day.
