Initial dataset upload (2026-07)

.gitattributes

@@ -58,3 +58,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+nerq-mcp-mortality-2026-07.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,97 @@
----
-license: cc-by-4.0
----
+# Nerq MCP-server repository mortality — dataset dump (2026-07)
+
+**Version:** 2026-07 · **As of:** 2026-07-22 · **Rows:** 28,824 · **License:** CC-BY 4.0
+
+© 2026 Nerq. Licensed CC-BY 4.0 — free to use, share, adapt, and use commercially, with attribution to
+Nerq (link back). No share-alike. Dated factual observations, without warranty. The continuously-updated
+series is available via the free Nerq API.
+
+## What this is
+
+A dated observation spine over the public GitHub repositories linked from indexed **Model Context
+Protocol (MCP)** servers, classified by repository liveness. One row per observation. **28,824 rows**
+in this dump, all from the seed census taken **2026-07-22**.
+
+This is the same population as **The Nerq Census No.1 / No.2 (2026)** ("The MCP Deadpool"), released
+as a dated observation spine rather than a one-verdict-per-repo summary.
+
+## Files
+
+- `nerq-mcp-mortality-2026-07.csv` — one row per observation, all columns below.
+- `nerq-mcp-mortality-2026-07.json` — the same rows as a JSON array under `rows`, preceded by a
+  metadata head (version, as-of, license, state counts, columns).
+
+## Schema (columns)
+
+| column | type | meaning |
+|---|---|---|
+| `entity_id` | text | GitHub `owner/repo` of the MCP server. |
+| `registry` | text | source registry tag. All rows in this dump: `mcp`. |
+| `observed_date` | date `YYYY-MM-DD` | the day this observation was taken (as-of). |
+| `state` | text | `deleted` \| `abandoned` \| `dormant` \| `alive` \| `unknown` (see thresholds). |
+| `http_status` | int / blank | raw HTTP seen (`200` / `404`), or blank if not read at seed. |
+| `last_commit` | timestamp / blank | the repo's own last public commit timestamp; blank if none read (e.g. `404`). Repository timestamp, NOT author identity. |
+| `source` | text | how the row was produced: `census-no2-seed` (all rows in this dump) or `re-probe` (future rows). |
+| `provenance` | text | `own` — every row is a Nerq first-hand read (the repo's own public `commits.atom` / HTTP status). None are ingested third-party data. |
+
+## State thresholds (exact)
+
+State is derived from `http_status` and the age of `last_commit` measured against `observed_date`:
+
+- `deleted` — HTTP `404` (repository deleted OR made private/inaccessible; code may live on privately).
+- `alive` — reachable (`200`) and last commit **≤ 180 days** before `observed_date`.
+- `dormant` — reachable and last commit **> 180 and ≤ 365 days** before `observed_date`.
+- `abandoned` — reachable and last commit **> 365 days** before `observed_date`.
+- `unknown` — state could not be resolved at seed (41 rows).
+
+Full derivation in `methodology.md`.
+
+## State counts in this dump (as of 2026-07-22)
+
+| state | rows |
+|---|---|
+| alive | 20,390 |
+| deleted | 3,448 |
+| abandoned | 2,686 |
+| dormant | 2,259 |
+| unknown | 41 |
+| **total** | **28,824** |
+
+## Interval-censoring (what "observed_date" does and does not tell you)
+
+Each row states that an entity was in a given state **on that observation date** — nothing more. A
+state transition (for example a repository going `alive → deleted`) can only be located to the
+**interval between two observations**, never to an exact instant.
+
+- This dump is a single-day seed (all rows `2026-07-22`), so for a `deleted` row the exact death
+  date is **unknowable** from this dump alone — it is censored to the interval between the repo's
+  `last_commit` and `2026-07-22`.
+- The living series adds `re-probe` rows over time, narrowing those intervals. Between probe passes
+  there are collection gaps; there is a **known collection gap on 2026-07-30** during which no fresh
+  observation was recorded, so any transition spanning that day is censored to the surrounding
+  interval rather than pinned to it. Gaps are disclosed, not smoothed over.
+
+## What we observe vs what we ingest
+
+- **Observe (this dump):** repository liveness — read first-hand from each repo's public
+  `commits.atom` feed and HTTP status (`provenance = own`). No API key, not rate-limited.
+- **Ingest (NOT in this dump):** upstream registry lifecycle fields (e.g. an MCP registry's own
+  `statusChangedAt`) are kept provenance-separate and are not mixed into these `own` rows.
+
+## Caveats
+
+- `404` = deleted **or** made private — we say "deleted / inaccessible", not "proven deleted".
+- `last_commit` is the default branch's last commit read from `commits.atom`; a small boundary class
+  near the 180-day line may differ from the GitHub REST API `pushed_at`.
+- This is a **prevalence** dump (state at the as-of date), not a lifespan study. MCP births are
+  left-censored, so median lifespan is not computable from this dump alone.
+
+## Cite
+
+Nerq. *MCP-server repository mortality — dataset dump (2026-07).* nerq.ai. CC-BY-4.0.
+
+**DOI.** This dataset is deposited on Zenodo with versioning. **Cite the Concept DOI**
+`10.5281/zenodo.XXXXXXX` (placeholder until minted) — it always resolves to the latest
+monthly version. For a reproducible snapshot, cite the version DOI of the release you used
+(e.g. the `2026-07` version). New versions are published **monthly**; corrections between
+releases appear first on the live series and are folded into the next monthly version.

methodology.md

@@ -0,0 +1,63 @@
+# Methodology — Nerq MCP-server repository mortality (2026-07)
+
+**As of:** 2026-07-22 · **License:** CC-BY 4.0 · Reproducible from public sources with no API key.
+
+This document states how each `state` in the dataset is derived so the dump can be independently
+reproduced and spot-checked.
+
+## 1. Population
+
+Distinct public GitHub repositories (`owner/repo`) linked from indexed Model Context Protocol (MCP)
+servers. 28,824 distinct repositories in this dump. Population definition is honest and bounded: it is
+the set of GitHub repositories reachable from the MCP servers Nerq indexes, not "all MCP servers that
+have ever existed."
+
+## 2. Reads (source of truth)
+
+For each repository, two first-hand reads are taken (`provenance = own`):
+
+1. **HTTP status** of the repository — `200` reachable, `404` gone (deleted or made private).
+2. **Last commit timestamp** — the latest entry in the repository's public `commits.atom` feed
+   (`https://github.com/{owner}/{repo}/commits.atom`). No API key. Not rate-limited. The timestamp is
+   the repository's own last public commit on its default branch; it is a repository fact, not an
+   assertion about any person.
+
+## 3. State derivation (exact rules)
+
+Let `age = observed_date − last_commit` in days. Then:
+
+```
+if http_status == 404:                      state = "deleted"     # gone or made private
+elif last_commit is unreadable/none:        state = "unknown"
+elif age <= 180:                            state = "alive"
+elif 180 <  age <= 365:                     state = "dormant"
+else (age > 365):                           state = "abandoned"
+```
+
+- **`deleted`** is authoritative on the HTTP read (`404`), heuristic on the cause (deleted vs private).
+- **`alive` / `dormant` / `abandoned`** are heuristic thresholds on commit recency. The thresholds
+  (180 / 365 days) are stated here and are the only knobs; changing them re-buckets the same reads.
+- **`unknown`** (41 rows) is recorded, not guessed — the row exists so the absence is auditable.
+
+## 4. Interval-censoring
+
+An observation dates a *state*, not a *transition*. A `deleted` row means "reachable-or-not was `404`
+on `observed_date`"; the actual deletion happened at some unknown point between the repo's
+`last_commit` and `observed_date`. This dump is a single seed day (`2026-07-22`), so death dates are
+interval-censored to `[last_commit, 2026-07-22]`. The living series narrows these with dated `re-probe`
+rows; collection gaps (a known one on **2026-07-30**) are disclosed, and any transition spanning a gap
+is censored to the surrounding interval rather than pinned.
+
+## 5. Reproduction
+
+To reproduce any single row: fetch `https://github.com/{entity_id}` (expect `200` or `404`) and
+`https://github.com/{entity_id}/commits.atom` (read the first `<updated>` timestamp), then apply the
+rules in §3 with `observed_date = 2026-07-22`. No credentials required.
+
+## 6. Known boundaries
+
+- `commits.atom` reflects the default branch only; a repo committing on a non-default branch could read
+  as more stale than it is.
+- The 180-day boundary class (repos whose last commit sits within a few days of the threshold) may flip
+  between `alive` and `dormant` depending on the exact read time.
+- `404` conflates deletion and privatization; the dataset does not claim to distinguish them.

nerq-mcp-mortality-2026-07.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0a0240b51f7b11a432aa8363055964d16ec07d727b40b703e0b920e16c0909a
+size 11827562