spb/datacenterindex

1# switch — Switch PRIME campuses23| | |4|---|---|5| Config | `config/connectors/switch.yaml` |6| Parser | `apps/worker/src/connectors/operators1/switch.ts` → `switch_campus_v1` |7| Kind / priority | operator / 1 |8| License / attribution | Publicly accessible operator pages; factual data only; © Switch, Inc. |9| Cadence | facility pages weekly · newsroom daily · index monthly · sitemap weekly |1011## What is crawled12- Yoast sitemaps `/page-sitemap.xml` and `/post-sitemap.xml` (148 posts) + RSS `/feed/`.13- Campus pages: `/core/` (The Core, Las Vegas), `/citadel/` (The Citadel, Tahoe Reno), `/keep/` (The Keep, Atlanta), `/pyramid/` (The Pyramid, Grand Rapids); `/rock/` (Austin) is matched if it appears. Market pages `/las-vegas/`, `/tahoe-reno/`, `/atlanta/`, `/grand-rapids/`, `/austin/` → `index`.14- Press releases are root-level WordPress posts (`/<slug>/`), so they are classified as `newsroom` at priority 40 from the post sitemap and the RSS; marketing pages (`/ebc/`, `/brochure/`, `/l/`, `/videos/`, careers/partnerportal/sustainable-data-centers two-level paths, …) are excluded explicitly and the GenericConnector relevance filter drops anything not about data centers.15- robots.txt has no disallows.1617## Access18Open site, L1 works (0 credits). `maxLevel: 4`, `maxCreditsPerRun: 30`.1920## Fields (method) — server-rendered facts only21The campus pages render their body client-side (the `<article>` is empty in the HTML), so only two server-rendered sources exist and are used:22- `<title>` / `og:title` "Switch – The Core Campus, Las Vegas, Nevada" → `name` "Switch The Core Campus", `city` Las Vegas, `regionName` NV, `countryIso2` US. When the title is a slogan instead (Pyramid: "GRAND RAPIDS: Largest, Most Advanced Campus in the Eastern U.S."), the campus/city/state come from the table of the four PRIME campuses published on Switch's own market pages (`fallback:switch-campus-table` in provenance).23- meta description "Switch's Tier 5 Platinum rated multi-tenant/colocation data center with up to 2.4 million sqft and 315 MW." → `tier` "Tier 5 Platinum", `plannedPowerMw` 315 ("up to" = design/full-build figure), `buildingSqm` 222,967 m² (2.4 M sq ft), `renewableClaim` ("The entire campus is powered by 100% green energy." on Pyramid), `description`.24- No address or coordinates on the pages → geo null. Certainty 0.7 (meta-only).25- Key `switch:<slug>`; `externalIds.switch_slug`.2627## Verification (2026-09-11)28- Discovery: 153 URLs → `facility_pages=4`, `newsroom=128`, `index=6`, default 15 (two-level marketing paths, now excluded).29- `--url` Core: Las Vegas, NV, US, Tier 5 Platinum, planned 315 MW, 222,967 m² — matches the live meta description.30- `--url` Pyramid: Grand Rapids, MI, US, planned 320 MW ("up to 320 megawatts"), 167,226 m² (1.8 M sq ft), "powered by 100% green energy".31- All 4 campus pages: 4/4 valid, 0 rejected; Citadel and Keep carry only name/location (their descriptions state no figures). Credits: 0.3233## Gaps / TODO34- Body content is JS-rendered; a Scrapfly-rendered fetch (30 credits) could expose per-campus specs — not worth it for four pages that change rarely. Revisit if Switch publishes spec pages server-side.35