Rohit Malekar

At GrantsScope, we’re pioneering the intersection of data science and AI to empower donors in the web3 public goods funding (PGF). Our mission? To revolutionize how donors identify and support grantees during PGF rounds. The Challenge:&nbsp;In large PGF rounds, where hundreds of grantees vie for attention, the brightest spotlight often falls on those with the strongest marketing. This imbalance creates a cycle where deserving projects may go unnoticed, leading to an uneven distribution of resources and opportunities. The Result:&nbsp;This disparity not only places additional burdens on grantees to amplify their visibility but also diverts their focus from creating meaningful impact. As a result, the true potential of PGF is left untapped, and innovation in public goods is stifled. Our Solution: GrantsScopeWe believe in leveling the playing field. GrantsScope harnesses the power of data analytics, machine learning, and advanced Large Language Models (LLMs) to enhance project visibility and democratize access to funding. Our experiments are designed to assist donors during quadratic funding rounds on Gitcoin Grants by<ul><li>Providing personalized recommendations, helping donors easily navigate the complexity of PGF rounds.</li><li>Offering self-discovery tools that empower donors to explore and connect with projects that resonate with their values.</li><li>Reducing information asymmetry, ensuring donors comprehensively understand the PGF landscape.</li></ul> Example #1: Use of RAG (Retrieval-Augment Generation) to allow donors to learn about grantees operating in causes they care about in GG19 <img src="https://giveth.mypinata.cloud/ipfs/Qmaj5hzYRVitzW5Bt71764X1WSrPbtZjCnSYfGAUPovWux" style="" width="563"> Example #2: Use of clustering techniques (HDBSCAN) to highlight summaries of related and distinct projects deployed during GreenPill x Octant Community Round <img src="https://giveth.mypinata.cloud/ipfs/QmUqVJfY7HWUov7ZwtqQqLXSVtDBBeXCQAbFbTdXsEKxoQ"> Example 3: Use of onchain donation history to drive personalized grantee recommendations for donors <img src="https://giveth.mypinata.cloud/ipfs/QmVe3K5sYCLDNptrq7yT552VedzJN3CqwgZsFPWU2ADUBK" style="cursor: nesw-resize;" width="563"> Our Impact: Visit <a href="https://grantsscope.xyz/" rel="noopener noreferrer" target="_blank">https://grantsscope.xyz/</a> to access apps, source code, and sample results.Join a growing community of over 500 unique users, as verified by StreamLit analytics, in exploring innovative applications that leverage conversational experiences using LLMs (Large Language Models) and clustering techniques. As the next milestone, GrantsScope will utilize existing onchain data and clustering techniques to offer personalized recommendations to donors in Gitcoin Grants (GG20) in April 2024. The Team: <a href="https://twitter.com/RohitMalekar" rel="noopener noreferrer" target="_blank">Rohit Malekar</a> currently maintains this project as a single-person team. It is built on data abstractions created by&nbsp;<a href="https://github.com/gitcoinco/grants-stack-indexer" rel="noopener noreferrer" target="_blank">Grants Stack Indexer&nbsp;</a>and&nbsp;<a href="https://davidgasquez.github.io/gitcoin-grants-data-portal/" rel="noopener noreferrer" target="_blank">Gitcoin Grants Data Portal</a> with plans to integrate with <a href="https://regendata.xyz/" rel="noopener noreferrer" target="_blank">Regen Data</a>.

At GrantsScope, we’re pioneering the intersection of data science and AI to
empower donors in the web3 public goods funding (PGF). Our mission? To
revolutionize how donors identify and support grantees during PGF rounds.
The Challenge: In large PGF rounds, where hundreds of grantees vie for
atten...

GrantsScope - Grantee Discovery Using LLMs