Technologies
Recent movement and use across the technology ecosystem.
What is moving from attention into code, adoption and operation — and where do channels diverge?
GitHub and Hacker News show attention/code; npm, PyPI, Hugging Face and marketplaces approximate adoption; Docker and cross-source relationships help distinguish experimentation from operational use. Units are never summed into a universal score.
Stars and weekly downloads come from the matched panel (the same entities in both weeks), not the whole ecosystem; each KPI has its own unit.
What grew or fell this week
Latest value per entity in the last 7 days against the same entity's latest value in the prior 7. Each row has its own unit and scale; do not compare bars across registries.
Which projects show up repeatedly in the radar?
Recurrence = days on which the project appears in the 90-day window (not the number of reads). Stars are only size context — a large project is not automatically the most recurrent.
Stars are shown only as size context; they are not used to manufacture a recurrence ranking.
Main usage and development signals
Preview of the top 4 of each ranking — code (GitHub), package adoption (npm, PyPI) and model use (Hugging Face). Each list has its own unit and criterion; nothing is summed into a score. "View all" opens the complete served list.
Criterion: current stars (size)
Criterion: days in the radar over 90d (not size)
Criterion: 30-day downloads (usage proxy)
Criterion: monitored-snapshot score
Criterion: monitored-snapshot score
How coverage is measuredRows written by collectors. Useful for auditing coverage and cadence; it is not ecosystem growth.
How much data we have from each source
Observations written by collectors and distinct entities tracked. This measures collection cadence and coverage, not ecosystem growth — real movement is in the matched panel above.
When interest, usage and development do not move togetherTechnology and geography cross-signals remain available on demand without dominating the primary read.
Technologies & Ecosystem Health
Four newly-connected sources, each with its own unit, window and maturity criterion. A reading appears here only once it clears its family's minimum history/coverage gate.
Homebrew installs over the last 30 days. Measures reach inside the macOS/dev ecosystem, not overall market share.
- Homebrew · installs (30d): 27 days of history · 10 tracked entities · latest reading: 21h ago · gate: ≥14d, ≥5 · reason: mature
Question volume and unanswered rate across tracked tags. Measures attention and friction, not adoption.
No mature metrics in this category yet.
- StackExchange · question volume: 11 days of history · 11 tracked entities · latest reading: 20h ago · gate: ≥14d, ≥3 · reason: insufficient history
- StackExchange · unanswered rate: 11 days of history · 11 tracked entities · latest reading: 20h ago · gate: ≥14d, ≥3 · reason: insufficient history
Average list price across tracked SKUs (AWS, Azure). Cost context, never a trend — no delta is computed here.
No mature metrics in this category yet.
- Azure · compute: 1 days of history · 1 tracked entities · latest reading: 634h ago · gate: ≥0d, ≥3 · reason: stale reading
- Azure · storage: 1 days of history · 1 tracked entities · latest reading: 634h ago · gate: ≥0d, ≥3 · reason: stale reading
- AWS · database: 2 days of history · 1 tracked entities · latest reading: 165h ago · gate: ≥0d, ≥3 · reason: insufficient coverage
- Azure · gpu: 1 days of history · 1 tracked entities · latest reading: 634h ago · gate: ≥0d, ≥3 · reason: stale reading
- AWS · compute: 2 days of history · 1 tracked entities · latest reading: 141h ago · gate: ≥0d, ≥3 · reason: insufficient coverage
- AWS · storage: 2 days of history · 1 tracked entities · latest reading: 597h ago · gate: ≥0d, ≥3 · reason: stale reading
- AWS · gpu: 2 days of history · 1 tracked entities · latest reading: 141h ago · gate: ≥0d, ≥3 · reason: insufficient coverage
- Azure · database: 1 days of history · 1 tracked entities · latest reading: 634h ago · gate: ≥0d, ≥3 · reason: stale reading
Cross-source relations across these four families: not yet enabled — no family has cleared a maturity review for a derived relation.
GitHub codeSize, recurrence and languages are complementary evidence. Open to compare rankings without treating them as the same metric.
GitHub · size and recurrence
Two distinct views: largest repositories by current stars and projects recurring most often in the 90-day window.
All present on 90/90 days: recurrence does not differentiate these projects; the order below follows stars.
Monitored software packagesnpm and PyPI have their own scales. Rankings remain available for investigation without dominating the primary read.
Monitored software packages
Ranking of the monitored snapshot; treat as presence signal, not market share.
Bar length is normalised separately inside npm and PyPI; do not compare absolute bar length across registries.
Hugging Face models30-day downloads are a usage proxy in the monitored corpus, not market share.
Hugging Face · observed use
Ordered by downloads over the last 30 days when available.
Community attentionHacker News measures discussion and editorial attention. Open to inspect stories; do not treat points/comments as adoption.
Hacker News · 30d
Stories deduplicated by URL/title. Editorial interest and discussion are not adoption.
Developer toolsMarketplaces show presence and accumulated use; the source does not always provide a comparable recent window.
Developer tools
Accumulated inventory of each collected marketplace: installs and rating are adoption context. Each marketplace has its own scale and ranking — do not compare figures across them.
Agenda and eventsThe full agenda remains available without competing with adoption and regime-change signals.
Technology events around the world
Full event explorer with filters, map and event names without semantic truncation.