ACM COMPASS is an international conference focused on the application of computing technologies for sustainability. COMPASS 2024 was held in New Delhi, from July 8th and 11th, and saw immense participation from researchers, practitioners, and policy makers, with over 270 attendees over four days.
Over the last fifteen years, members of the COMPASS community have impacted millions of people through their work by founding startups and social enterprises based on their research, making policy interventions, and participating in large development projects.
With a view to scale the impact further, a new initiative introduced at COMPASS this year was the Research to Impact Collaboratives (RICs) concept. 8 RICs were formed through pre-consultations between researchers, practitioner organizations, policy makers, and community stakeholders, to identify specific research agendas that can be solved and taken to scale rapidly. These RICs spanned areas of building data exchange networks of air pollution sensors, innovations for accessibility using smartphones and LLMs, digital public infrastructures of datasets and algorithms for climate change adaptation, use of AI advances in agriculture and health, and building new methods for skilling at population scale, among others. COMPASS aims to facilitate the formation of similar RICs in other countries, networking across RICs, and guide and nurture the RICs to flourish and create change.
The CoRE stack network also ran a RIC that brought together over 40 people, 14 thought provoking thoughts, and 4 group activities to work out important next steps. A brief description of the discussions is below.
Vision of the CoRE stack: Aaditeshwar Seth presented the vision of networked co-creation that the CoRE stack wants to enable, highlighted the relevance of data and algorithm standards for multiple research and development groups to contribute and use one another’s outputs, and of being community driven from the outset in adopting use-cases for collective pursuit.
Agroforestry and biodiversity: ISB (Abhijeet Parmar) presented their work on using photographs of tree barks and leaves to identify the tree species, an effort that can speed up conducting tree censuses in forest plots. SNU (Snehasis Mukherjee), SIMATS, and WASSAN presented methods using computer vision techniques on drone imagery to assess the health of tree plantations. ATREE (Ashish Kumar) and IITD presented their ideas on building an end-to-end feasibility site assessment method for tree plantations, based on the climatic, terrain, and water requirements of different tree species.
Agriculture: Plaksha University (Anupam Sobti) presented their work on using satellite and drone imagery to identify sugarcane plantations, and to monitor the application of different nutrients to sugarcane plots. WASSAN (Prachi Patil) presented an exciting new way to understand rainfed agriculture by monitoring soil moisture; this can lead to new ideas to revive traditional inter-cropping systems that are resilient to drought. SocialWell (Atanu Garai) presented their work on forecasting of climate data, that can be used to assess the risk of different kinds of cropping patterns.
Data and algorithm standards: To enable networked co-creation, three presentations highlighted different elements of important next steps. R. Nandakumar (formerly at ISRO) presented the need to build interaction standards on data objects, so that feedback, linkages, and navigation, can be more streamlined. WELL Labs (Craig Dsouza) updated about several global initiatives in progress to create metadata standards for geospatial datasets, and how to think about measuring the impact of putting new standards in place. The University of Cambridge (Anil Madhavapeddy) presented their work on a planetary computing infrastructure to keep track of different base data versions that are used in the computation of downstream data products.
Community engagement: FES (Himani Sharma, Chiranjit Guha) presented their work on crowd-sourced data collection of well measurements and other ecological indicators, and the challenges therein to do this at scale. Auroville Consulting (Athithiyan MR.) presented their Life Lands (LiLa) platform to assess land suitability for restoration activities. Gram Vaani (Aila Dutt) presented their ideas on creating community engagement with data, starting first with building a system-based understanding of social-ecological systems, and then to think in terms of visualizations and whether they are being effective in conveying new insights.
The agenda document has links to these talks.
This was followed by four group activities. The data and algorithms standards group led by Craig Dsouza (WELL Labs) went further by taking the example of Land Use and Land Cover (LULC) datasets to lay out a pathway of what specifications need to be put into place. The productization group led by Ajay Tannirkulam (Magasool) took the example of farm-level insurance products to lay out a roadmap of new datasets and tools that are required to build open-source tools to guide farmers to make insurance claims. The community engagement group led by Aila Dutt (Gram Vaani) brought new ideas and guidelines to keep in mind while building out engagement around data. The collaboration group led by Deborah Thomas (Living Landscapes) discussed essential ingredients to make collaboration successful, such as taking ownership, being transparent, listening actively to all participants, and to build a democratic governance structure.
If you are interested in joining the network then do drop an email to [email protected]
Also check out this blog post [link CoRE Stack Geospatial Layers and Datasets – corestack] from a few months ago about various novel geospatial data layers that have been generated so far. Let us know if you need access to such data for your watershed development or ecological restoration activities.