Methodology
How the data behind this site was built
Every number on this site comes from one pipeline: raw pages in, structured remote positions out, then deduplication, normalization and aggregation. This section documents each step and its limits.
Source pages
975
saved as plain text
Raw positions
4,512
extracted as remote
Unique positions
4,022
after deduplication
Skill-tagged
1,265
name at least one skill
The pipeline
| Step | Stage | Result |
|---|---|---|
| 1 | Collect | 975 source pages saved as text |
| 2 | Extract | 4,512 remote positions found by claude-opus-5-5 |
| 3 | Deduplicate | 4,022 unique positions by title and company |
| 4 | Normalize | 1,556 distinct skills after aliases and variant merging |
| 5 | Aggregate | Counts by skill, family, level and pair |
| 6 | Analyze | Categories, insights and ranked recommendations |
Details
Methodology
Data collection
What the 975 source pages were and how they were gathered.
MethodologyExtraction of remote positions
How a language model turned raw pages into 4,512 structured positions.
MethodologyDeduplication
How 4,512 raw positions became 4,022 unique positions.
MethodologySkill normalization
Aliases, naming variants and how each skill gets one name.
MethodologyRole families and seniority
The nine families and seven levels every position was assigned to.
MethodologyLimitations
What this data can and cannot tell you.
MethodologyTwo runs compared
A consistency check between the July and October 2026 extraction runs.