Data Playbooks

From messy upload to finished deliverable: quality audits, cleanups, exploratory analysis, styled Excel workbooks (including a saved template you already branded), interactive dashboards, living Google Sheets, and reviews that run on a schedule. Attach a file or connect a source, then ask in plain language.

CSV Quality Audit

• Good for: unknown or messy datasets
• Best setup: uploaded CSV plus code execution
• Example prompt:

CSV audit

Inspect the attached CSV and give me a data-quality audit. Flag missing values, inconsistent formatting, duplicate records, suspicious columns, and the highest-impact cleanup steps before analysis.

Spreadsheet Normalization

• Good for: turning operational data into analysis-ready tables
• Best setup: spreadsheet upload plus explicit schema expectations
• Example prompt:

Normalize spreadsheet

Clean and normalize this spreadsheet into an analysis-ready table. Standardize dates, currencies, and categorical values. Return both the cleaned output and a short log of the transformations you applied.

Exploratory Data Analysis

• Good for: first-pass understanding of a dataset
• Best setup: CSV or workbook plus code execution
• Example prompt:

EDA

Run exploratory analysis on the attached dataset. Summarize distributions, outliers, correlations, and the most decision-relevant patterns. Keep the write-up business-readable, not notebook-jargon heavy.

KPI Anomaly Scan

• Good for: recurring operational reviews
• Best setup: structured KPI table and a defined review cadence
• Example prompt:

KPI anomaly scan

Scan this KPI dataset for anomalies, breaks in trend, and unusual combinations of signals. Prioritize findings by business importance and suggest the three follow-up questions I should investigate next.

Structured Workbooks Without Formula Pain (Excel)

• Good for: workbooks people actually open — vendor evaluations, project trackers, models, board reports
• Best setup: name the shape you want and hand over the content; Alfrada OS handles layout, styling, and formulas. If you already saved a matching .xlsx under Settings → Templates, that file is used instead of a generic preset.
• Example prompts:

Vendor comparison workbook

Build me a vendor comparison workbook for Acme, Bolt, and Corvid: sections for security, pricing, and support with weighted criteria in each, and a summary that ranks the vendors by weighted score. I want to adjust the weights in Excel and watch the ranking update.

Project tracker with dropdowns

Set up a project tracker for the website relaunch: columns for task, owner, status, priority, and due date. Make status and priority dropdown menus so nobody can free-type them, and freeze the header row.

Alfrada's Excel tool comes with six ready-made workbook shapes — questionnaire, tracker, comparison matrix, financial model, KPI dashboard, and executive report. Ask for one by name and you get a styled .xlsx with live formulas, not a static table: numeric strings become numbers, formula text becomes a real formula (a missing = on MAX(...) is rescued), and changing an input cell in Excel recalculates everything downstream. Long columns can be auto-fitted; a logo or chart image can be embedded in the sheet.

Fill The Workbook Template You Already Branded

• Good for: next quarter's model, next client's scorecard — the same furniture, new numbers
• Best setup: save the .xlsx once under Settings → Templates, then name that template in the ask
• Example prompt:

Fill a saved workbook template

Use our saved financial-model template. Fill it for Meridian Logistics with these assumptions, don't restyle the sheets, and check the layout before you hand it over.

Inspect copies the stored template into this session (the original is never edited), fills every , reports anything still showing , and can visually check for clipped columns or leftover placeholders. Release note: Workbooks And Word Files Now Fill Like The File You Sent.

Edit The Spreadsheet You Already Have

• Good for: fixing and extending an existing workbook without rebuilding it
• Best setup: upload the .xlsx and describe the changes; you get the same file back, edited in place — inspect first, then targeted writes, not a new workbook
• Example prompt:

Workbook write-back

Open my uploaded workbook, fix the Region column values — standardize them to North, South, East, West — add a formula column for gross margin, bold the header row, and give the file back.

From Anomaly Scan To KPI Dashboard (HTML Dashboard)

• Good for: turning the KPI Anomaly Scan above into something you can hover, filter, and share
• Best setup: run the scan first, then ask for the metrics-wall template
• Example prompt:

Anomaly dashboard

Take the anomaly scan you just ran and build a metrics-wall dashboard of the affected KPIs — metric cards with deltas and sparklines, plus a bar chart of the biggest movers — and let me export it as a PDF.

HTML Dashboard ships six starter templates; the metrics-wall template is a responsive grid of metric cards with deltas and sparklines, made for exactly this kind of readout. For the full tour — including dashboards that refresh themselves — see the Dashboard Playbooks.

My GSheets Workflows

Connect your Google account under Settings → Tools & Data → Data Sources → My GSheets and Alfrada OS works your real spreadsheets in place — reading, appending, deduping, and charting without a download-edit-reupload loop.

• Good for: living trackers and shared sheets that other people also touch
• Best setup: name the spreadsheet and tab, and be explicit about the key column when deduping
• Example prompts:

Weekly tracker update

Append this week's numbers to my Growth Tracker sheet, dedupe on the week-start column, and refresh the trend chart on the Summary tab.

Clean a shared sheet

In the shared Leads sheet, find the rows where the email column is blank or malformed, fix what you can from the company column, and flag the rest in a new 'Needs review' column.

My GA4 Weekly Report

Connect your Google Analytics property under Settings → Tools & Data → Data Sources → My GA4 and get trend reads without opening the GA4 console.

• Good for: a weekly pulse on traffic and conversions, in plain language
• Best setup: a connected GA4 property, plus Task Scheduler for the recurrence
• Example prompt:

GA4 weekly trend read

Every Monday pull last week's sessions, conversions, and top pages from GA4 and give me a one-page trend read against the week before.

The "every Monday" part becomes an automation: Alfrada OS sets it up with Task Scheduler, and each run lands in your chat list as its own conversation. Manage it under Settings → Automations — see how automations work.

Recurring Data Review (Task Scheduler)

• Good for: making the KPI Anomaly Scan above happen without you remembering it
• Best setup: a source Alfrada OS can reach on its own each run — a connected sheet or GA4 property, not a one-off upload
• Example prompt:

Monthly anomaly review

Set up a monthly task for the 1st at 9am: pull the latest numbers from my KPI tracker sheet, run an anomaly scan on them, and leave me a prioritized findings list with the three follow-up questions I should ask.

Ship The Findings (Doc Patch / Report Generator)

• Good for: the last mile — turning the analysis into a document someone signs off on
• Best setup: tell Alfrada OS the format — Report Generator produces polished PDF and Word reports; Doc Patch edits the Word file you already have without breaking its formatting
• Example prompts:

Analysis to PDF report

Write the analysis up as a polished PDF report: executive summary, key charts, findings ranked by business impact, and an appendix with the cleanup log.

Patch a Word report

Patch the updated figures into my uploaded Word report — replace the numbers in the Results section and the summary table — without touching the document's formatting.

Under the hood — tool ids, presets, and templates

• excel_ops (Excel) — saved-template fill via inspect on templates/<file>.xlsx then find_replace + optional visual_check; structured workbooks via create_structured_workbook with preset ids questionnaire, tracker, comparison_matrix, financial_model, dashboard (the KPI one), and executive_report; write-back via write_data_to_excel, apply_formula, autofit_columns, insert_image, format_range, and add_data_validation.
• html_dashboard (HTML Dashboard) — template ids blank, executive_summary, kpi_grid (the metrics wall), research_report, comparison, timeline.
• gsheets_* (My GSheets) — 17 tools, including gsheets_append_values, gsheets_upsert_rows, gsheets_lookup_row, gsheets_create_chart.
• ganalytics_* (My GA4) — 36 tools, including ganalytics_run_report, ganalytics_run_realtime_report, ganalytics_run_funnel_report.
• task_scheduler (Task Scheduler) — recurring background runs.
• doc_ops (Doc Patch) — surgical .docx edits that preserve formatting.
• report_generator (Report Generator) — PDF and Docx output.