Detect Errors and Bias Before Reporting

Today's Goals

1. Find duplicate IDs and invalid numeric values without altering evidence.
2. Separate clean-baseline totals from contaminated-file observations.
3. Rewrite an AI summary to state denominator, scope, and uncertainty.

Plain-Language Walkthrough

A clean-looking total can contain two different defects

Today the lab's registration CSV is intentionally contaminated. It has ten data rows rather than the clean baseline's eight. One additional row repeats EntryID A06 with the same Night Market seat and USD 5 receipt. Another row, A09, gives Garden Walk a Seats value of the word two, not the number 2. Both rows are visible on purpose. The exercise is not to guess the correct repair and make the file tidy. It is to detect that the file cannot be used as an authoritative replacement without a source-owner decision.

A duplicate identifier is not automatically a harmless duplicate line. The second A06 might be a copy error, a separate booking incorrectly assigned the same ID, or a revision that should supersede the first. Those possibilities lead to different totals. Naively including it adds one Night Market seat and USD 5 to the clean numbers. Quietly deleting one row asserts that the entries are identical in meaning. Preserve both original lines, mark the collision, and record who must decide the authoritative record. Until then, publish the reviewed eight-row baseline if a report is required, with a clear statement that the contaminated file is excluded.

Sign in and link WeChat to read

Sign in, then link your WeChat mini program account on the account page. Linked accounts share one free article per day. Unlock additional articles with rewarded videos in the mini program.

Scan in WeChat to open Dayward AI. Open the same course and chapter, choose the rewarded video, then refresh this page after completion.

Unlocked — refresh to readSign in / scan with WeChat →