Before building anything I wanted to know what this dataset can and cannot support. Phase 0 was about scope: reading the Launch Library 2 API, sketching the model, and writing the decisions down.
The grain is one row per launch, keyed on the launch UUID. Five nested objects come out of the API cleanly: provider, rocket configuration, mission, pad, and status. Which of these become conformed dimensions, and which end up as fact attributes, is a Phase 3 decision. Status in particular is more likely a degenerate attribute than a dimension. Two relationships are many-to-many (program and configuration families), so they are left for a later version.
The decisions that were not obvious:
- Vehicle identity. A "rocket" in the API is an instance; what I want to compare is the configuration. So vehicle identity is rocket.configuration.id, and configurations get their own load.
- Outcomes and censoring. There are nine statuses. They collapse into four real outcomes and five schedule states. That split matters for any later reliability analysis, so I keep it explicit instead of flattening everything to success or failure.
- Partial failures. Partial failure stays as its own outcome rather than being forced into success or failure, and the reasoning is in the scope document.
- Dates. net is the launch date, and its precision varies for older records, so I anchor on net rather than a window start.
- Orbital scope. I scope to orbital launches using mission.orbit, and Unknown is its own category instead of being assumed orbital. Suborbital is excluded.
Then I checked the most recent 100 launches for missing data:
Every record had a mission and an orbit, so on this sample mission.orbit is a clean filter. The orbit split: 60 Low Earth, 11 Sun-Synchronous, 10 Polar, 7 Geostationary Transfer, 2 Elliptical, and 1 Medium Earth, plus 4 suborbital and 5 Unknown.
One caveat: these are the most recent 100 launches, the best case. Records from the 1960s and 70s will have more gaps, so I will re-run the same check against the full history in Phase 1.
The output of this phase is a written scope document in the repo: the grain, which statuses count as success, how partial failures are handled and why, how the orbital scope is drawn, and what I am deliberately not modeling in v1.
Next: Phase 1, the full backfill from Launch Library 2 into bronze.