Builder segments
Builders are split into two primary subgroups: developers and researchers. Developers are blockchain accounts with a github repository and working code contributions. Researchers are blockchain accounts with a Discourse profile.
Developers and researchers are further segmented by a variety of characteristics. The developer segment is primarily characterized by software languages.
Repo segmentation criteria:
- Categorize repos by language, where:
- Repo is not a token repo (e.g., /assets, /tokens, token-list, etc.)
- Languages is not equal to JavaScript, TypeScript, CSS, HTML
- Developer type is not equal to bot
- The total bytes written for the language is greater than or equal to 25% of all Non-front-end language bytes written
- Project name is not equal to Gitcoin Grants (Gitcoin Grants repo lists the repos of all grantee recipients)
- Ranking projects within the language segment
- Normalize the bytes written, language percentage share of the project, stargaze count, contributor count, repo count, and fork count across all projects in the resulting cohort. We use a simple scale between 0-1.
- Assign weights to each normalized category, as follows:
- bytes written: 15%
- language dominance: 10%
- stargaze count: 25%
- contributor count: 20%
- repo count: 5%
- fork count: 25%
- Rank using final weighted score
- Filter for top 250 projects
- Assigning categories: Top Project, Leader, In-The-Mix, Laggard:
- Calculate quartiles for top 250 projects by weighted score
- Assign labels: Top Project = top quartile; Leader = 3rd quartile; In-The-Mix = 2nd quartile; Laggard = bottom quartile