Portfolio Projects
Intermediate 12 min read Portfolio Projects

Presentation and Documentation

Make the work survive contact with an audience

Presentations and docs are the delivery layer of an analyst's work. A tight 10-minute review and a README that stands alone make your analysis reproducible, reviewable, and promotable.

presentationdocumentationREADMEdeckreproducibilityhandoff
Prerequisites: Data Storytelling
Before we start

What you will learn

  • Structure a 10-minute analytics review
  • Write a README that reproduces the work from scratch
  • Design the handoff so next week's analyst inherits the context
Introduction

What is this?

Presentation and documentation cover both the live review and the durable artifact. A great review is 10 tight minutes: context, one insight, evidence, recommendation, next steps. Great documentation — a README with data dictionary, method, reproduction steps, and known limitations — lets the work be trusted and continued without you.

Motivation

Why do you need it?

  • A review wins the decision; the docs protect the decision afterward.
  • Reproducible work is checkable work — reviewers trust what they can rerun.
  • Handoff quality determines whether your insight becomes a product or a footnote.
Think about it

Mental model

Imagine

A pilot's cockpit checklist

Think of

The presentation + docs pair

The live checklist gets the flight done; the manuals keep every future flight possible. Reviews decide, docs endure.

Mechanics

How it works

  • Build the deck around one decision: context, insight, evidence, action.
  • Rehearse to a timer and cut anything that does not serve the takeaway.
  • Write the README from a blank folder: can a stranger rerun it?
  • Archive the method, data dictionary, and limitations beside the code.

Live review wins the decision; docs keep it trustworthy after.

Review
Evidence
README
Handoff
Real world

In production

  • A monthly growth review ends with one ask: which experiment to fund.
  • A portfolio README includes the environment and data seed script.
  • An analyst offboard passes a documented repo, not a chat thread.
Guidelines

Best practices

  • Rehearse out loud and keep the review under its timebox.
  • Put reproduction steps and the data dictionary in the README.
  • List known limitations explicitly — they build more trust than polish.
Watch out

Common mistakes

  • A 40-slide deck whose insight is on slide 28.
  • A README that cannot be followed from an empty checkout.
  • Docs that describe what was done but never why.
When things break

Debugging tips

  • If rehearsal runs long every time, cut a section, then cut parts of another.
  • If a reviewer cannot rerun the work, add a `make setup` section or pinned env file.
Interviews

Interview checkpoint

Test yourself

Quick quiz

0/2
1

A strong analytics review centers on

2

A README that lets a stranger rerun the work requires

Go deeper

Official documentation