Technology / a one-minute field guide

Package Pulse

How do TypeScript package downloads vary by day?

heatmapchartflipCardlistaccordionwebViewstatusLight
Daily package downloads2026-08-08 / Day 1: 221955152026-08-08 / 2: 223655252026-08-08 / 3: 435085942026-08-08 / 4: 02026-08-08 / 5: 453171382026-08-08 / 6: 455852262026-08-08 / 7: 02026-08-15 / Day 1: 236279002026-08-15 / 2: 229883152026-08-15 / 3: 414990042026-08-15 / 4: 465465802026-08-15 / 5: 454750802026-08-15 / 6: 460840462026-08-15 / 7: 429551762026-08-22 / Day 1: 237379762026-08-22 / 2: 233086152026-08-22 / 3: 451502232026-08-22 / 4: 475810632026-08-22 / 5: 467779182026-08-22 / 6: 464012552026-08-22 / 7: 412411602026-08-29 / Day 1: 229689832026-08-29 / 2: 233169632026-08-29 / 3: 440198192026-08-29 / 4: 501193272026-08-29 / 5: 515131232026-08-29 / 6: 02026-08-29 / 7: 46015098

Data illustration, not an app screenshot. Snapshot retrieved 2026-09-07T16:27:03+00:00. Open the template for refreshed data.

App link not opening? Your installed version may not recognize website template links yet. Use Templates inside CAR-TER, or download the file and open it with CAR-TER from Files or the share sheet. Get CAR-TER on the App Store.

In the app: Templates → Featured → Package Pulse. The app catalog version is signed and unlocks the template’s controls; manual file imports may require Control Pack for some controls.

Try it in 60 seconds

  1. Read the question. Check the source response and retrieval time before trusting any reading.
  2. Compare the color pattern, then open Explore for exact values.
  3. Tap either reading card to reveal what it means. Open Explore for another view and the source records.
  4. Find the observation time. Downloads are not unique users, active projects, quality, or security. Includes automated installs.

The idea worth keeping

A daily series reveals a rhythm; its total needs the exact reporting window.

A short script you can read aloud

“How do TypeScript package downloads vary by day? This layout starts with that question, not a wall of controls. The data comes from npm registry. Compare the color pattern, then open Explore for exact values. The reading cards tell you what to notice; turn one over when you need context. A daily series reveals a rhythm; its total needs the exact reporting window. One important limit: Downloads are not unique users, active projects, quality, or security. Includes automated installs.”

Suggested animation scene

Start with one labeled observation. Arrange the observations into cells; fade in the color scale only after the audience knows what one cell means.

Storyboard: 0–8 seconds, ask the question. 8–25 seconds, introduce one observation. 25–45 seconds, perform the gesture above. 45–60 seconds, show the timestamp and limitation. Keep labels readable, offer a reduced-motion cut, and use actual source values or a clearly marked replay.

How this becomes a layout

The provider publishes data → a small read-only adapter selects and names the useful fields → CAR-TER controls bind to those fields. All controls in this template share one endpoint. Polling is cached for 3600 seconds; that is a retrieval interval, not a promise that the source updates that often.

Remix exercise

Duplicate this layout. Keep the source and units unchanged, but replace the primary visualization with the alternative in Explore, or move the record list above it. Ask a new reader which arrangement answers the opening question faster. Changing location or query parameters requires a new allowlisted server-side source, not an unwired selector.

Source, rights and limitations

npm registry · Exact upstream query · Usage terms / attribution

Public npm download-count API; service terms apply. No package source or release artifacts copied.

Downloads are not unique users, active projects, quality, or security. Includes automated installs.

This is a selected result window, not a full archive. Missing numbers are not replaced with zero. Lines connect available observations; consult records for gaps. A green READY badge means the last retrieval succeeded, not that the underlying observation is recent. STALE means a retained prior response; UNAVAILABLE means no usable cached response. If your device goes offline, check the displayed retrieval time: the last screen can remain visible.

Open-data and publicly accessible API are different promises. This collection labels each source’s conditions; it does not assert that every provider’s database or API is open-source software.