EFCD profiles a data feed and compares it with a baseline. It looks for changes in the schema and field statistics.
Behavior
Inputs can be HTTP JSON, local JSONL, or CSV. Commands build profiles, run comparisons, and render HTML reports. Optional webhooks send the resulting alerts.
A baseline can use a previous profile, a pinned reference, a rolling window, or a seasonal policy. Those choices determine what counts as a change.
Scope
The implementation is a scaffold for developing feed-monitoring workflows. A changed field statistic is an observation to investigate; its significance depends on the feed and the chosen baseline.