Happy holidays

I muse about my latest holiday project - explaining why some primary schools are not popular schools.

Happy holidays
Photo by Kevin Woblick / Unsplash

For readers in the US, today is 4 July, so there's something about Independence day. Back in Singapore, it's Youth Day, so if you find lots of kids hanging out in malls, you tell yourself they earned it this time.

Holidays are essential for your well-being. While I was a litigation associate, I planned my holidays six months ahead and protected them with my life. It wasn't enough. Going overseas helps you relax, but it doesn't give you the space to check in with yourself. Or explore something new.

And this post is about exploring something new.

Get some reading done.

I have bought several books over the year. The Employment Law book I acquired with my Academy of Law credits is still in its shrink wrap. A holiday seems like a good time to get it done.

There is one book I would like to spend time on Hands-on Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems.

Having some experience with Natural Language Processing, I would like to broaden my knowledge. Rather than working with text, this book would hopefully increase my ability to work with numbers. If you don't already know, in Natural Language Processing, the text is transformed into numbers or vectors and then processed by machines.

Data Science with Judgement Data – My PDPC Decisions Journey
An interesting experiment to apply what I learnt in Data Science to the area of law.
You can check out some of my excursions here.

It's time to get back to the source.

The Maths is making me dizzy... but to be frank, it is more readable than an ordinary computer science textbook.

Let's hope I am still awake after all these equations.

A very long holiday project

Along the same lines as reading something different from your daily work or life, I firmly believe you should work on an off-kilter and exciting project. It helps you explore what you learned differently and very often pays dividends for your day job.

The curious incident of playing 5 million Monopoly JR games with Python
My lockdown project goes on its own sojourn through Python multiprocessing programming with interesting, unexpected results.
I tried to justify my losing streak in a children's board game by repeatedly making a computer play 5 million games. It turned out to be an experiment on high-performance python programming.

I've been wondering what I should do next, and I think I have a great idea.

2021 was the first time I went through the process of entering my eldest daughter into Primary 1. We decided early that we would not go for the rat race. We're not enamoured by "top" or "popular" schools. There might have been a morally justifiable reason for our choice, but I reckon we didn't think the stress was worth it.

Despite consciously avoiding the "rat race", it turned out to be futile. We went into the lottery for a place in an "unpopular" school and had to wait a week to learn of our daughter's fate. Luckily, this story ended OK.

Commentary: P1 registration changes – what would it take for us to believe ‘every school is a good school’?
The unnecessary stress over Primary 1 registration can be mitigated if we resolved to make “every school is a good school” a reality, says June Yong.
This article nicely summarises what all parents might have to go through to enrol their children in Primary 1.

Based on my experience, at least, while emphasising "popular" school is understandable, I realised I knew very little about the registration situation. Most existing studies focused on "popularity", which isn't so helpful.

In ST's study, you know what schools are "popular", but you have no idea why.

So, I think there's some scope for machine learning. For example, does the age of an estate affect the competition for places? Are schools with keen competition for places located in particular areas of Singapore? Does school affiliation or religious affiliation affect competition?

My son registers for Primary 1 in 3 years. That's more than enough time to build a machine model to predict how much stress I have to go through.


I'm no neophyte. This can be a challenging project which will summon every fibre of my being -- from finding datasets, transforming them and visualising them. Hopefully, I'll get to document them in this blog for your reading pleasure.

And yet, it's a holiday project, so if I get bored or fail, I will move on. (I don't expect my son to hate me for it either.)