Reporting analysts & coordinators
Use the AI assistant your employer allows to list, check and draft around the weekly report: the export you did not make, measures that need a written meaning, the value that stands out, the summary people read instead of the tables. You find every number in the file, and definitions, personal data and what leaves the company stay with their owners.
Get startedNo card needed to begin · I already study here
9 short lessons · 56 min of audio · a final exam and a course certificate
Use an AI assistant to clean an export, write down what each measure means, question a value that stands out, draft the weekly summary and caption a chart, while every number is found in the file by you, every measure keeps its owner, and personal data stays where your employer's policy allows.
fixes kept apart from decisions, the columns nobody asked for, and the list of what this file cannot tell us: The export you did not make; The columns you did not ask for; What the file can answer
one written meaning and one owner for every word you count, the outlier asked why before it is written up, and every number found in the file: What on time means; Ask why before you write; Every number from the file
the line under each chart that says what it leaves out, a dashboard described for people who never open it, and the reporting card for the busy week: What the chart leaves out; For the people who never open the dashboard; The reporting card
A reporting card on one page, kept in your work tools: the untouched export and the cleaning log, what the file cannot tell you, the glossary with owners, every number found in the file, a line under each chart on what it leaves out, the note for people who never open the dashboard, and a hands-off list with real names.
The certificate names that artefact and the tools you used — not a title. It carries a code anyone can check: verify a certificate.
The opening of lesson one, exactly as it plays inside — and the exercise that comes with it, which you can do right now without an account.
Most reports start with a file somebody else made. You ask a system for its data, and it hands you an export: a download of rows and columns, often a spreadsheet.
It is rarely tidy. The header row appears twice. A totals line at the bottom looks like one more row of data. Some dates are written day first, some month first. The same depot is spelled two ways. There are rows for test parcels, and rows that look exactly like the one above.
You will meet Ines throughout this course. She is a reporting analyst at a regional parcel delivery company. Every week she turns exports from several systems into one report for the operations director and the depot managers. The systems team that runs the delivery system is what this course calls the data owner: the people responsible for a system's data and for what it means.
Here is the trap. Ask an assistant to "clean this", and you may get back a tidy table that looks ready. Along the way it may have removed what it took for duplicates, filled in blanks, merged the two spellings, and dropped the totals line, or kept it. Some of that may be right. But the table no longer says what was done to it, and a report built on it is one you cannot explain.
A fix keeps what every value means. A decision changes what a value means, or whether a row counts. Write both down, and never let a decision pass as a fix.
Try it now · Split an export's problems into fixes and decisions
Where:Your approved work assistant, on a copy without the columns you do not use
Here is this week's export. Do not change anything yet. List every problem you can see, and for each one say whether fixing it keeps the meaning, or changes what a value means or which rows count. Do not remove or fill in any rows.
The rest of the catalogue is there when you need it — but you do not have to choose anything today. One course, one short lesson at a time, one page at the end.
Get started