Housing supply analysis

From the change wiki
Jump to navigation Jump to search

This is a general guide to analyzing the housing situation in any part of the world.

The goal is to answer some basic questions:

  • Is the housing stock sufficient to meet the needs of the people?
  • If not, what are the most efficient ways to fix it?

First thing you need is census data. This comes in different formats depending on which city/country you want to analyze.

Intended results

Simple mode: Existence of housing

  • 2 stacked bar charts.
    • One for housing units, and the other for population units.
    • Visually compare to see if in principle there are enough homes suited for each demographic.

Advanced mode: Distribution of housing:

  • Full table (will be described below)

Data categorization


A population unit is just a corny way to say "a person, or a group of people who probably intend to live together".

Population unit categories:

  • s0: single adult with no kids
  • c0: couple with no kids
  • s1: single parent with 1 kid
  • c1: couple with 1 kid
  • s2: single parent with 2 kids
  • c2: couple with 2 kids
  • s3: single parent with 3 kids
  • c3: couple with 3 kids
  • : ...and so on and so forth...

Of course there are a lot of annoying nuances within this. How do you define kid and parent? How do you define single and couple? What about familes who don't want to live together? What about friends who do? The world doesn't fit neatly into boxes.

No matter how we define the categories, it won't be perfect - there will always be a few people misrepresented. But that's ok - we're just trying to get a general sense of the demand for housing.


For this, we don't care so much about the type (apartment, condo, townhouse, single-detached house, etc). We just look at the number of bedrooms.

Housing unit categories:

  • st: studio or bachelor
  • 1b: one bedroom
  • 2b: two bedrooms
  • 3b: three bedrooms
  • : ...and so on and so forth...


Census data might, for example, only have a column for '1 bedroom', '2 bedrooms', and '3 or more bedrooms'. In that case, you can use the notation '1b', '2b', and '3+', and call it a day. Same for population/singles/couples/families.

Bar chart

Generate 2 stacked bar charts that look like this:


The width of each bar in the first chart should be the number of housing units.
The width of each bar in the second chart should be the number of population units.
The visual scale (units per pixel) should be the same for both.

Full table

Table should look something like this

. hl st 1b 2b 3b 4+

The 'hl' stands for 'homeless'.
The 'va' stands for 'vacant'.
Each square should say the number of population units living in each type of dwelling.


It might be very hard to find the census data needed to fill this table.

Also, it isn't clear how to document, for example, 3 single adults living in a 3-bedroom apartment.

Alternative: Just make a table of number of bedrooms with number of people living there?

Join the discussion to try and find a better way to analyze the distribution of housing.


Coming soon.

See also