HARBOR | Usability Test Coding Sheet

Here is the finished, editable workbook — one XLSX, built with xlsxwriter and read back with openpyxl.

What is inside

• Observation detail — 60 rows (P01–P12 × T1–T5), fields Participant, Task, Completion status, Time (seconds), Error count, Issue tag, Observation notes. Frozen panes at A4, autofilter A3:G63, six data-validation rules: dropdowns for participant, task, status (4 options) and issue tag (6 options), plus 0–3600 s and 0–20 error bounds. Status values are colour-coded.
• Coding dictionary — defines the four completion statuses and six issue tags (what to look for, typical severity, how to code), a task reference with scenario prompts, and the coding conventions. Its counts are live COUNTIF / AVERAGEIFS formulas.
• Task summary — headline metrics, the per-task table using COUNTIFS for completion and AVERAGEIFS for average time and errors, an overall row, a status mix, and three charts (completion rate, average time, status doughnut) in deep-sea blue, sky blue and warm white.
• Findings and next steps — five sample findings ranked by severity × frequency (27, 12, 10, 8, 3), each with evidence, a recommended next step and an explicit "Yes — needs validation in real testing" line.

Every sheet carries the banner "Simulated data · Not real participants", the HARBOR name is fictional, and no real names or personal data appear anywhere.

Verification (122 checks, all passed)

• Container: zip integrity clean, charts present in the package, opens in openpyxl in both formula and data_only modes.
• Data: exactly 60 rows, no blanks, 5 rows per participant, 12 per task, all statuses/tags within the dictionaries.
• Formulas: 91 formula cells (COUNTIFS, AVERAGEIFS, COUNTIF, IFERROR, priority score), every one carrying a cached value so the preview shows the KPIs without recalculation — overall completion 85.0%, average time 124.65 s, average errors 1.47.
• Independent recomputation: per-task attempts, completed, completion rate, average time, average errors, abandoned and the status mix were recalculated from the raw 60 rows and matched the cached values exactly (tolerance 1e-9). No error tokens, no external links.
• Validation/charts: 6 data validations confirmed with their sqref ranges and option counts; 3 charts confirmed with their category and value ranges.
• Layout: explicit widths on every used column of every sheet, plus a width/height audit for clipping (no unwrapped overflow; only short section and total rows are under 30 pt by design).

Limits to note

• No renderer or office suite was installed, so I could not produce a pixel rendering; visual coverage is the deterministic width/row-height audit plus the checks above, not a rendered preview.
• The file name uses a hyphen ("HARBOR - Usability Test Coding Sheet.xlsx") because "|" is invalid in Windows filenames; the exact title "HARBOR | Usability Test Coding Sheet" is preserved in the banner of every sheet and in the document properties.
• The findings are hypotheses from constructed records, not measured prevalence.