Checklist

Predictive Maintenance: Data Requirements & Labeling Checklist (Pilot Readiness)

Interactive checklist to verify sensor, event, labeling, and data-pipeline readiness for condition-based or predictive maintenance pilots. Includes verification tips, common fixes, and structured fields to record findings, showstoppers, and an overall readiness score.

Interactive Tool

Predictive Maintenance: Data Requirements & Labeling Checklist (Pilot Readiness)

Use this checklist to confirm that the data and labeling foundation for a predictive-maintenance pilot is reliable and actionable. Each item includes a short verification tip and common fixes to help teams move from 'data exists' to 'data is usable for modeling.' Save findings so they can be tracked across iterations.

Name this pilot, proof-of-concept, or model so records are traceable.
Person completing the checklist.
Use YYYY-MM-DD or the date format your team prefers.
Verification tip: Confirm a stable unique identifier for each monitored asset and a consistent hierarchy (site>plant>line>machine>component). Common fixes: reconcile naming maps, add canonical ID mapping table, enforce tagging at source.
Verification tip: Check that timestamps across sensors, PLCs, and logs share a common timezone and reference. Common fixes: apply consistent timezone, correct offsets, or add ingest-stage alignment.
Verification tip: Measure uptime, dropouts, and error codes. Common fixes: replace faulty sensors, fix wiring, add redundant channels, or filter noisy readings in preprocessing.
Verification tip: Ensure sampling rate captures the dynamics of the failure mode (e.g., vibration signatures require higher sample rates than slow temperature drift). Common fixes: increase sample rate for critical signals, apply feature extraction for low-rate signals.
Enter typical sample rate or interval. Leave blank if not applicable.
Verification tip: Confirm labeled failure events exist and are timestamped. Common fixes: align event windows, enrich logs with failure annotations, and capture near-miss or precursor labels.
Count of labeled events (approx). Helpful for planning modeling approach (supervised vs anomaly detection).
Consider consistency, timestamp accuracy, and label granularity.
1.0 10.0
Verification tip: Can you link failure or repair records to asset IDs and timestamps? Common fixes: add repair identifiers, standardize log fields, or automate log ingestion.
Verification tip: Verify retention period, raw vs aggregated storage, and file formats. Common fixes: adjust retention, implement tiered storage, or standardize file schemas (Parquet/CSV/TSDB).
Select any applicable considerations that affect access, sharing, or labeling work.
Estimate the work to annotate events or confirm labels if automatic labels are insufficient.
Describe who will label, tools (e.g., labeling UI), sampling strategy, and quality checks.
Note key preprocessing steps (resampling, filtering, FFTs, feature windows, outlier handling).
List any issues that must be resolved before modeling can proceed (e.g., no failure labels, inaccessible logs).
Record prioritized remediation actions, owners, and target dates.
Use this to quickly gauge pilot readiness and whether the project should use supervised learning, semi-supervised approaches, or anomaly detection.
1.0 10.0
You can explore this tool now. Sign in or create an account to save your responses and return to them later.
Sign In to Save Create an Account
Make this tool part of your work

Save a personal copy, bring it to your team, or tailor the questions and workflow to fit what you are hungry to improve.

Member customization and team collaboration are coming soon.
Discuss Ask the Librarian

Discussion

Comments and conversation will live here.