A free Singapore public housing historical price research tool

I created a historical HDB price research tool for anyone who needs it

Background

Remember the Singapore public housing price analysis I did previously? It was based on the publicly available dataset provided by the Singapore government. While I plan to do more analyses on this public data, I figured I can build a proof-of-concept Singapore public housing resale price research tool for anyone who needs it, since I also already built a proof-of-concept tool to check the availability of my bookmarked NLB books.

Why reinvent the wheel?

In case anyone is asking, the Singapore public housing agency, Housing Development Board (HDB) does provide historical public housing transaction information here.

Fig 1 : The search interface of the HDB portal

The search results does provide quite a bit of information.

Fig 2 : The search results — which is pretty awesome!

And this includes some nearby amenities of the public homes transacted.

Fig 3 : Amenities near the homes sold

Unfortunately, the housing data provided by HDB feels like a data dump that doesn’t enable further insights or actions. And it doesn’t help that the website is a bit slow to load, and requires users to click through a few buttons for each request they want.

Who is this tool for?

The aim of my tool is to give people a quick self-serve to understand the historical public housing prices of certain regions in Singapore. I am assuming this would be both potential buyers and sellers of Singapore public housing. As I don’t want to overload the server of the free hosting site that I am using, I am only showing transactions from since 2022. I also feel that showing recent transactions would be more useful for users.

Finally, let me explain briefly how to use my tool

How to use the tool

1. Filters

To allow users to select the housing transactions that they want, we definitely need filters, and they appear on the left side of my main dashboard page. To ensure that the user doesn’t not try to load too much data, the user can only select one town at a time.

Fig 4 : You don’t need to put in the entire Street Name to do a search

2. Summary

One thing that I really wanted is this summary view based on flat models. Other than Total Transactions, the other variables of Remaining Lease, Resale Price, Flat Area (Sqm) are average values. To allow a quick comparison, I also added the total and average of all transactions for all filtered results.

Fig 5 : A quick summary of Total Transactions and average Remaining Lease, Resale Price and Sqm.

3. Resale Price Distribution

Next, I included my favourite visualisation, the box plot. This box plot allows me to look at how the resale prices are like for different flat model types beyond their average and median values. The general idea of each bar is that it shows the maximum, minimum, 3rd and 1st quantile, and median value of each category.

For example, just looking at the visualisation for Ang Mo Kio public housing transactions, New Generation flat models seem to costs the least overall as a model type, while Improved flat models have quite a huge variation of prices, from around 200K to 1.2 million SGD!

Fig 6 : Improved flat models in AMK have such huge differences in price!

4. Transactions table

Naturally, I am including each transaction and it’s related information. This isn’t very different from what HDB is providing.

Fig 7 : Plan, simple, number of transactions

5. Map view

I included a map visualisation to show all the transactions that the user have filtered for. For this, I need the user to click on the “Show Map” checkbox to show the visualisation up after the user has added the criteria he wants for his property search. I hope that this will allow the user to do a bit more surface investigation on the neighbouring amenities.

However, I have to admit that the map experience isn’t as great as what you can get from Google Maps. Also, because I am hosting this tool on a free service, loading too many transactions onto the map may cause the map to crash and disappear. If this happens, just untick and tick the “Show Map” checker option again after you have reduced the number of transactions that you are investigating.

As always, “Done is better than Perfect”.

Caveats

• As I am putting (hosting) the tool on a free service (called Streamlit), the tool may be wonky at times if too many people is using it. However, I am not even sure if anyone will use this tool, so this for now is both my least of concerns and also the first concern I like to highlight.
• I am updating the data manually, so the data may not always be very updated. Especially if no one cares about it (see point 1). That said, I have included a date to show when the data is updated.
• Housing coordinates depend on OneMap API, which may not always return results.
• Lastly, (see point 1) if anyone finds this tool valuable or have any feedback on this, feel free to contact me on Linkedin!

Final words

While the data is publicly available, it is kept in a format that may be difficult for non-technical people to use. Nonetheless, everyone needs housing, and I hope that this tool that I built can help fellow Singaporean residents with their home sale and purchase decisions. Housing is a very sensitive issue for many Singaporeans, and I hope I can do my little part to add a little bit more transparency to the public housing market, and help make it more efficient.

And finally, here is the link to my web tool that I built!