Priority: P3
Owner: —
Why it's parked
Visual sugar with low insight value and high complexity: the layout depends on the chosen metric, so each metric combination would need its own city. Interactive version rejected.
What survives
- A video rendering (FFmpeg, like Gource) is acceptable for marketing — computing for an hour is fine offline. Technique: start from the final state to freeze layout and districts, then reveal buildings as you move forward in time
- The higher-value variant is a diff between two releases: which concepts / buildings appeared, how much extra code to hold in your head. Combine with the LLM ("what did the agent add these last days?"). Consider this a separate P1 later.
Source: CodeGraph backlog extracted from the 2026-09-21 weekly (David / Jérémie).
Priority: P3
Owner: —
Why it's parked
Visual sugar with low insight value and high complexity: the layout depends on the chosen metric, so each metric combination would need its own city. Interactive version rejected.
What survives
Source: CodeGraph backlog extracted from the 2026-09-21 weekly (David / Jérémie).