Instrument the study table and plot

Instrument the study-hours table plus hours-vs-pass plot so a run records enough structure to debug offline: counts, latency if relevant, pass/fail of mean hours, max hours, and plot file exists with xlabel/ylabel set, and a stable stage name. Redact secrets and raw credentials from every event.

import json
print(json.dumps({'rows':4,'mean_hours':5.25,'max_hours':12,'plot':'hours_vs_pass.png'}))

Expected evidence: summary metrics JSON. Prefer JSON or structured text you can grep in CI over prose logs for data-tables-and-plots.

Turn one historical failure—especially wrong column names—into a test that fails if the signal disappears for the study table and plot. Observability without a failing test is optional decoration; observability with a test is part of the data-tables-and-plots artifact.

Draft a three-field event for the study table and plot: stage, ok, and one domain field derived from mean hours, max hours, and plot file exists with xlabel/ylabel set. Add fixture_id or docs_version when content can change. Explicitly list fields that must never appear (tokens, passwords, raw prompts) because publishing learner names from the CSV in a shared plot title is in scope for this lab.

Wire one assertion that fails if the study table and plot event is missing after a run. Observability that cannot fail a test will not survive contact with a busy data-tables-and-plots repository.

Imagine a teammate opens only your event stream after a bad deploy. Could they tell whether four-row study.csv with hours and passed columns was wrong, whether wrong column names, or a plot that saves without axis labels returned, or whether publishing learner names from the CSV in a shared plot title slipped through? If not, rename fields until those three stories are distinguishable.

At the testing and observability stage for data-tables-and-plots, the job is narrower than finishing a product demo. You are creating one progressive evidence piece about four-row study.csv with hours and passed columns that later pages inherit without redefining success. Keep that fixture small enough to inspect by hand, keep outputs copy-pasteable as text, and refuse to narrate this baseline as if it were a production SLA: hand-computed mean of the four-row fixture.

For this page specifically, success looks like a structured event schema locked by a test while still centering the user decision to inspect typical hours, an extreme value, and the hours/pass relationship. If you cannot point to a file, command, or assertion that proves that for the study table and plot, stay on this page instead of advancing.

Pandas glossary

Previous · Next