My Hobbies

From vintage mechanics to modern DevOps — here’s what I enjoy building and maintaining.


Our Love for Red

1977 Classic Mini 1100 Special

Our love for the UK goes beyond whisky and rolling hills — it extends to Sir Alec Issigonis’ ingenious design of the Mini. What started as a clever solution to post‑war fuel shortages became one of the most iconic small cars ever built. We’re fortunate to own a 1977 Classic Mini 1100 Special, a special Belgium edition.

When we first got the car, it had been neglected for years. Maintenance was patchy, it was colored fifty shades of black/purple (even some parts with glitter), and the electrics were unreliable. Since then, we’ve been steadily bringing it back to life: repainting panels, fixing rust within our possibilities, fixing wiring gremlins, replacing windows, and tackling countless small jobs. Some repairs we handle ourselves — like the above and normal maintenance — while bigger projects are entrusted to a local garage. It’s a balance of DIY passion and professional support, each step making the car a little better than before.

The Mini has a personality that makes it impossible not to smile, every time we open the garage door. Driving it is pure joy: the direct steering, the mechanical feedback, the sense of connection to the road.

1977 Classic Mini 1100 Special
Our 1977 Mini 1100 Special.

2004 Ducati ST3

The Ducati ST3 is a sport‑tourer that blends Italian character with practical long‑distance ability. It has the unmistakable exotic quirks that Ducati is known for: a desmodromic engine, revving up to 11,000 RPM, sharp handling, a very satisfying exhaute note... And it doesn't like low revs, or speeds below 30 km/h. That's 'character' for you.

Much of the maintenance I handle myself: valve clearance checks, timing belts, fluid changes, brake pads, and chain care. These jobs demand patience and a steep learning curve, but they make the bike feel more personal and rewarding. For safety checks and maintenance beyond my skills, I rely on a trusted garage.

Every ride reminds me why I keep investing time in it. The sound of the engine, the way it pulls through corners, and the confidence it inspires on longer trips all make the effort worthwhile. It's only been in my pessession since the end of 2024, so I'm looking forward to many more miles of adventures ahead.

Ducati ST3
My Ducati ST3 sport‑tourer.

News (NOS) Sentiment Lab

I run a small Airflow pipeline that scrapes NOS headlines, classifies sentiment with Ollama, and stores the results in PostgreSQL for this site to consume.

The sentiment dashboard below shows the overall sentiment and category breakdown for today, the last 7 days, and the last 30 days. It's a fun way to see how news sentiment fluctuates over time.

The scores are normalized between -1 (very negative) and 1 (very positive), with 0 being neutral. The labels are derived from the average score: negative, neutral, or positive.

Today overall

NEUTRAL

Score: -0.2

49 articles

7 day average

NEGATIVE

Score: -0.4

244 articles

30 day average

NEGATIVE

Score: -0.3

403 articles

Today by category

Category
Score
Articles
  • BINNENLAND
    -0.6
    9
  • BUITENLAND
    -0.5
    10
  • ECONOMIE
    -0.5
    10
  • POLITIEK
    -0.3
    10
  • SPORT
    0.6
    10

7 day category average

Category
Score
Articles
  • BINNENLAND
    -0.6
    56
  • BUITENLAND
    -0.6
    68
  • ECONOMIE
    -0.4
    29
  • POLITIEK
    -0.3
    21
  • SPORT
    0.1
    68
  • VOORPAGINA
    -1.0
    2

30 day category average

Category
Score
Articles
  • BINNENLAND
    -0.6
    94
  • BUITENLAND
    -0.7
    116
  • ECONOMIE
    -0.4
    41
  • POLITIEK
    -0.4
    28
  • SPORT
    0.2
    118
  • VOORPAGINA
    -0.7
    6

QuickDraw Playground

Draw a simple sketch and send it to my QuickDraw classifier. The current model knows three classes: bicycle, cat, and coffee cup.

The browser converts your drawing to a 28×28 grayscale matrix before it is posted to the prediction service running in the ML platform.

Draw something, then ask the model for a prediction.

White strokes on a dark background match the training data best.


This portfolio website: nanderspeerstra.com

Automated pipeline from commit to live site — my garage for code, just with fewer oil stains.

Delivery flow

Code starts in GitHub, moves through automated release and CI steps, is packaged in Docker, deployed on microk8s, and served to the public through Cloudflare Tunnel.

🚀 GitHub 🤖 Release Please 🔄 Automerge 🛠 GitHub Actions 🐳 DockerHub 🖥 microk8s 🌐 Cloudflare Tunnel ✨ nanderspeerstra.com

Application

Django powers the site and renders the pages visitors see.

Data

PostgreSQL stores portfolio content and sentiment data for the dashboard.

Automation

Airflow keeps the news sentiment pipeline running in the background.