FluxFae/Synthetic_Clinical_Telemetry_and_Operator_Tickets

Creator Dataset — Synthetic Clinical Telemetry and Operator Tickets

FluxFae/Synthetic_Clinical_Telemetry_and_Operator_Tickets · dataset version 1.0.0 · generated 2026-07-01T00:00:00Z

A small, fully synthetic dataset for early-warning scoring on wearable patient telemetry, plus a ticket corpus of operator notes for text classification. Every row is produced by a deterministic generator (seed=42): no real patients, no real tickets, no PHI, no scraped data.

What this dataset is not

• No real patients. No record here describes a real person, and no value was sampled from a clinical source.
• Not for clinical use. Nothing here is a medical device, a validated risk model, or clinical guidance. Do not deploy any model trained on it.
• Not a question-answering set. The ticket table is a text classification task: a note in, a category out. It contains no question/answer pairs.

Files

Every file is published unchanged from the generator's export. These are the exact digests the export's own manifest.json records, so you can verify a download byte for byte:

file	bytes	sha256
audit_traceability_log.csv	15430	c29c09e335a248545fdfc24a2a480a1de1669c5cba6eaa5837b1e7cbb01bf1c1
centro_servizi_tickets.jsonl	326887	e1e6993db202ff0ae01ea04573bf7397ad5623410e0f1bf37c3b5f3e77f78c88
data_dictionary.md	17305	1be8b7f66ea6a1e516781d68fc034884ca20955a987ddff9c38c0e94fdc72afb
patients.csv	12837	4e4b3ad7632c4f9aea3ed1dcde3365167546e26bdff6427ccff809553b066b80
telemetry_edge_cases.csv	7687	df2225cec69f44ea0240e0abe42e43fae707bc275561ef10a8fca98b6bc0825a
telemetry_vitals.csv	1096989	4e347f2d08b49e9a9ab7a554a937d222c49f293f19abc8caeb8ba1795e34455a

manifest.json and validation_report.json are included in data/ as provenance: the manifest declares row counts, column lists, the seed and the model used for the free-text column, and the validation report lists the 14 internal consistency checks that this export passes.

Tables

table	rows	columns	file
patients	500	5	patients.csv
telemetry_vitals	6000	22	telemetry_vitals.csv
telemetry_edge_cases	50	11	telemetry_edge_cases.csv
centro_servizi_tickets	500	10	centro_servizi_tickets.jsonl
audit_traceability_log	50	15	audit_traceability_log.csv

tickets.parquet at the repository root holds the same 500 tickets as data/centro_servizi_tickets.jsonl, split into one Parquet file so load_dataset works in one line. Both representations carry identical values.

The ticket split

The 500 tickets are disjoint and frozen:

• 350 train — review_status: not_required. Intended for training, few-shot prompts, or historical context.
• 150 golden — review_status: pending_review. The held-out evaluation set for a classifier or retrieval system that has to separate simple technical faults from hidden clinical alarms.

Honest disclosure: the golden rows are labeled by the generator, not confirmed by a human annotator. review_status: pending_review is carried in the data on purpose and is not a placeholder to be ignored. Treat those 150 labels as a synthetic reference, and report your own agreement rate against them rather than assuming they are ground truth.

category takes eight values: device_alert, connectivity, battery, calibration, software, user_error, alarm, other. priority takes low, medium, high.

How to load it

from datasets import load_dataset

# Point at the Parquet file explicitly: the repository root also holds data/*.csv, so an
# implicit scan would leave the loader guessing which files form the dataset.
tickets = load_dataset("FluxFae/Synthetic_Clinical_Telemetry_and_Operator_Tickets", data_files="tickets.parquet", split="train")
print(tickets.features)
print(tickets[0])

Raw artifacts, for consumers that want the files exactly as exported:

from huggingface_hub import hf_hub_download

patients_csv = hf_hub_download("FluxFae/Synthetic_Clinical_Telemetry_and_Operator_Tickets", "data/patients.csv", repo_type="dataset")

The column semantics — type, role (id, time, feature, target, protected, metadata, split), range, allowed values and derivation — live in data/data_dictionary.md. Read it before using any column: several columns are targets and must never be used as model inputs, and one column is a protected attribute intended only for fairness checks.

Provenance

field	value
dataset version	1.0.0
seed	42
generation timestamp	2026-07-01T00:00:00Z (fixed, never the wall clock)
free-text model	qwen3:1.7b, temperature 0, thinking disabled
prompt version	notes_prompt_v1_en
determinism	two independent runs produce byte-identical files

operator_notes is the only generated free text; it was produced once and cached, and the cache is the reproducibility contract. Every other column is arithmetic on the seed.

Intended use

• Practising early-warning-score pipelines and deterioration-label logic on tabular telemetry.
• Benchmarking text classifiers or RAG retrieval over synthetic operator notes.
• Teaching data validation: one table ships deliberately malformed raw values (telemetry_edge_cases.csv) for a validator to catch.
• Testing data contracts, schema checks and reproducibility tooling.

Out of scope: clinical decision support, publication of clinical findings, any claim about real patient populations, and any fairness conclusion about real people.

License

Released under CC BY 4.0. You may use, share and adapt the data, including commercially, as long as you give appropriate credit.

Citation

@dataset{creator_dataset_v1.0.0,
  title  = {Creator Dataset: Synthetic Clinical Telemetry and Operator Tickets},
  year   = {2026},
  note   = {Synthetic dataset, CC BY 4.0},
  url    = {https://huggingface.co/datasets/FluxFae/Synthetic_Clinical_Telemetry_and_Operator_Tickets}
}