Turn the business question into a project plan

Understand the method

A useful portfolio project shows how you work, not only what software you can name. This capstone brings together the questions, preparation, calculations, exploration and communication developed in Modules 1–5. The fictional Northstar operations lead asks whether May’s larger completed-order value supports expanding an initiative. Your task is to produce a reviewable answer and identify the next evidence needed. It is acceptable for that answer to be “investigate further,” provided the reasoning is clear.

Define a bounded question

Use the supplied April–May 2026 extract. It contains order identifiers, order dates, channels, statuses, order values in EUR and recorded delivery durations. There are no customers, advertising costs, margins or experiment assignments. Frame the project around the observed completed-order change, channel composition and data coverage. Do not promise a profitability model or causal evaluation that the inputs cannot support.

The source is deliberately familiar: the raw and verified CSVs are unchanged from Module 4, and the dashboard reference is the worked example from Module 5. Integration is the new challenge. You must explain why each step is necessary, connect the outputs and show which work is your own. Reusing a teaching case is legitimate when its origin and your contribution are visible.

Define the finished submission

Plan an unchanged raw file, a prepared copy, a correction log, a metric contract, an editable analysis and a dashboard or clearly labelled static views. Add a concise case study, a decision note and a verification record. Keep the folder small enough that a reviewer can find the source, method and main conclusion without opening a dozen unexplained screenshots.

Set acceptance criteria before calculating. Required figures must reconcile to the completed-order population. Segment totals must agree with the monthly baseline. Missing delivery measurements must remain distinguishable from zero. The narrative must match the selected dashboard view. Every headline figure should be traceable to a source field and an explicit method. These are concrete criteria; “looks professional” alone cannot establish analytical quality.

Plan the sequence

Use eight milestones: scope, preparation, reproducible metrics, exploration, dashboard, written explanation, evidence packaging and independent extension. Save intermediate work at each milestone so a later chart can be traced to its inputs. Choose a primary calculation route: a spreadsheet using the methods from Module 2, or local SQLite using the supplied setup and starters. You may compare both, but you do not need to create unnecessary parallel systems.

State the project identity honestly

Label the final work “fictional guided portfolio project using DataCareerHub learning data.” Do not describe Northstar as a real client, invent a role you held or claim that a recommendation generated savings. A strong submission can describe an actual analytical contribution: resolving source issues, documenting denominator choices, testing interactions and explaining uncertainty. The final lesson adds a clearly separated independent extension so your own decisions are easier to recognise.