gh-star-region-breakdown

Generate a region breakdown chart for: $ARGUMENTS

Steps

1. Fetch stargazer data

Run the CLI to fetch all stargazers with locations (this caches the data):

npx gh-star-region <owner/repo> --no-open

This will paginate through all stargazers via the GitHub GraphQL API and cache the results.

2. Find unclassified locations

Run the unclassified locations script:

npx gh-star-unclassified <owner/repo> > /tmp/unclassified.csv

If all locations are already classified, skip to step 5.

3. Classify locations into regions

Split

/tmp/unclassified.csv

into batches of ~200 and spawn parallel subagents to classify each batch. Each subagent should:

• Read its batch file (e.g.,

  /tmp/locations_batch_0.csv

  )
• Determine the region for each location
• Write output to

  /tmp/locations_classified_0.csv

  with columns:

  location, count, region

Classification rules:

• Use standard region names (e.g., "South Korea", "United States", "China")
• US cities/states -> "United States"
• Chinese characters (深圳, 北京, etc.) -> "China"
• Korean characters (서울, etc.) -> "South Korea"
• Ambiguous/joke locations ("Earth", "localhost", "Matrix") -> "Unknown"
• Keep the exact original location string

4. Update location map

Merge the new classified locations into

~/.gh-star-history/location_region_map.json

, adding any new mappings that don't already exist (skip "Unknown" entries).

5. Generate the chart

Run the CLI again to generate the chart with updated mappings:

npx gh-star-region <owner/repo>

This will open the chart in the browser automatically. Do NOT run a separate

open

command.