The Hiring Landscape Has Changed

• Everyone has access to AI tools now
• Technical skills alone don’t differentiate you
• Employers want to see how you think, not just what you can code
• A portfolio shows your judgment, communication, and process

What AI Can’t Show For You

Your Portfolio Is Your Proof of Work

A portfolio isn’t a gallery of finished projects — it’s evidence of how you think through problems.

• It complements (not replaces) your resume
• Gives interviewers something concrete to discuss
• Shows you can communicate technical work to non-technical audiences
• Demonstrates you’re someone who ships — not just someone who studies

What Makes a Great Analytics Portfolio?

1. A clear “About Me” — who you are, what you’re interested in
2. 2-3 well-documented projects — quality over quantity
3. Process, not just results — show your reasoning
4. Clean presentation — design matters, even for data people
5. Easy to find — a real URL, linked from your resume and LinkedIn

Quick Tour: sepd.dev

🔗 sepd.dev

• Built entirely with Quarto
• Hosted free on GitHub Pages
• Blog posts, CV, about page
• Each post is a .qmd file — code + narrative

Other Portfolio Examples for Inspiration

Take a look at these for ideas:

• 🔗 qmdclub.com — a gallery of Quarto sites
• 🔗 Awesome Quarto — curated list of Quarto projects
• 🔗 posit::conf(2025) Workshop Materials — “Branded Websites, Presentations, Dashboards, and PDFs with Quarto”

What We’re Building Today

By the end of this section, you’ll have:

• ✅ A GitHub repository for your portfolio
• ✅ A live Quarto website on GitHub Pages
• ✅ A development environment with AI assistance built in
• ✅ Auto-deployment via GitHub Actions (no manual branch wrangling)
• ✅ Your first project post scaffolded and ready to write

Step 1: Create Your Repo from the Template

1. Make sure you’re logged in to github.com
2. Navigate to the template repo:

🔗 github.com/phdjsep/gh-pages-quarto-portfolio-template

3. Click “Use this template” → “Create new repo”
4. Name it: <your-username>.github.io
5. Click “Create Repository”

Step 2: Launch a Codespace

1. From your new repo, click the green “Code” button
2. Select the “Codespaces” tab
3. Click “Create codespace on main”

Step 3: Preview Your Site

Once your Codespace is ready, open the terminal and run:

quarto preview

Enable GitHub Pages

After pushing the Actions workflow:

1. Go to your repo → Settings → Pages
2. Under “Source”, select gh-pages branch
3. Click Save

Step 5: Create a Project Post

In your terminal:

mkdir -p posts/my-first-project
touch posts/my-first-project/index.qmd

Then add this YAML header to index.qmd:

---
title: "Predicting Customer Churn with Logistic Regression"
author: "Your Name"
date: today
categories: [machine-learning, python, classification]
description: >
  An end-to-end analysis exploring what drives customer churn
  using a telecom dataset, with actionable recommendations.
---

AI Is Your Portfolio Co-Pilot

AI tools can help you:

• Scaffold content — first drafts of your about page, project descriptions
• Write & debug code — Copilot autocompletes in your Codespace
• Troubleshoot Quarto — ask an AI when rendering fails
• Polish writing — improve clarity and conciseness
• Brainstorm — what projects should you showcase?

Demo: AI-Assisted Content Generation

🎯 Goal: Use AI to draft an “About Me” section, then refine it

When to Use AI (and When Not To)

Anatomy of a Great Portfolio Project

Every project post should answer four questions:

1. What problem did you solve? ← This is where most people fail. Lead with the “so what?”
2. How did you approach it? ← Methods, tools, data sources
3. What did you find? ← Key results, visualizations, insights
4. What did you learn? ← Reflections, limitations, next steps

Structure Your Post

## The Problem

Customer churn costs telecom companies billions annually.
I explored a dataset of 7,000 customers to identify the
key drivers of churn and build a predictive model.

## My Approach

- Exploratory data analysis to understand feature distributions
- Feature engineering: tenure buckets, monthly charge ratios
- Logistic regression (chosen for interpretability)
- Evaluation: confusion matrix, ROC-AUC

## Key Findings

[Include a visualization here]

The strongest predictors of churn were...

## Reflections & Next Steps

If I were to extend this analysis, I would...
The main limitation was...

Adding Code & Visualizations

Quarto renders code blocks natively. Include your actual analysis:

```{python}
import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv("data/churn.csv")
churn_by_tenure = df.groupby("tenure_bucket")["churn"].mean()

fig, ax = plt.subplots(figsize=(8, 4))
churn_by_tenure.plot(kind="bar", ax=ax, color="#4A90D9")
ax.set_title("Churn Rate by Customer Tenure")
ax.set_ylabel("Churn Rate")
plt.tight_layout()
plt.show()
```

AI Prompt for a Project Write-Up

Try this prompt pattern with Claude or ChatGPT:

Your Deployment Pipeline

With GitHub Actions set up, your workflow is now:

graph LR
    A[Edit .qmd files] --> B[Git commit & push]
    B --> C[GitHub Actions builds site]
    C --> D[Live on GitHub Pages]
    style A fill:#4A90D9,color:#fff
    style D fill:#2ECC71,color:#fff

Checklist Before You Leave Today

• Repository created from template (<username>.github.io)
• Codespace launched and Quarto working
• Site previewed and customized
• GitHub Actions workflow added
• GitHub Pages enabled from gh-pages branch
• First project post created
• About page drafted (AI-assisted is fine!)

Level Up: What to Do Next

• This week: Publish your first project post
• This month: Add 1-2 more projects from your coursework
• Customize: Explore themes, add a CV page, get a custom domain
• Share: Add the URL to your LinkedIn, resume, and email signature
• Maintain: A portfolio that hasn’t been updated in a year sends the wrong signal

Resources