Glint Analytics
We exist to empower and reward analysts for their brilliance through a tokenized economy for Blockchain Intelligence.
Demo
About
The problem it solves Current blockchain analytics solutions are centralized and extractive, failing to fairly incentivize analyst contributions. Analysts often work for free or recognition, leading to burnout and decreased output as rewards diminish. On Glint, dashboards are minted as NFTs on Base, with all interactions recorded through weighted voting. As our build progresses to production release, these votes will convert into rewards based on the value each analyst's creation provides to the community, as determined by the community themselves. A Glint dashboard NFT effectively serves as the creator's intellectual property, recorded immutably on Base. This not only showcases their work but also enables value sharing among all network participants, including creators, promoters, curators, and stakers. By leveraging Base in this way, we are building a community-driven platform that actively rewards proactive analytics and encourages the exploration of macro trends, deep dives into protocol sectors, and analysis of emerging blockchains, driving the future of blockchain intelligence.
How it's made
Challenges we ran into We encountered significant difficulties in describing our data schema to ChatGPT, which resulted in SQL code that did not meet our quality expectations. This issue highlighted the need for more effective communication of our schema details. To address this, we provided ChatGPT with comprehensive information about our tables, including detailed descriptions of their relationships and specific column details. This helped improve the context ChatGPT had when generating SQL code. Additionally, we identified that certain column names were causing confusion, so we renamed these columns to be more intuitive and descriptive. This renaming process significantly improved ChatGPT's understanding of our schema, leading to more accurate SQL code generation. We also experimented with different versions of ChatGPT to achieve reliable and consistent results. This experimentation allowed us to identify the version that best understood our requirements and provided the highest quality SQL code. Another improvement we made was incorporating timestamps into our data. By clearly defining the format and purpose of these timestamps, ChatGPT was able to interpret and handle them more effectively, enhancing the accuracy of temporal queries and operations in the generated SQL code.