Housing and Population Analysis
So far this data exploration has focused on vacant properties and fire code examinations. This is only part of the picture. A city isn't just the buildings in it but also the people. I wanted to start looking at how many people are in Saint Paul, has it been increasing, decreasing, staying level? How does that compare to the housing available? Do we have an overabundance of housing, a shortage? lets see if we can find out.
Population
Let's start by looking at the people in the city. For this I found data from the Metropolitan City Council, after a recommendation from another local in area. The Metropolitan City Council is a regional policy making body for the twin cities. The population data available from Metropolitan City Council has population data from the Census Bureau and the Metropolitan Council Research accumulated between the years of 1970 to 2024. On exploration of the data I found the Census Bureau ranges from 1970 to 2020 in 10 year segments. In addition to this data the Metropolitan Council Research added data from 2021 through 2024.
You may notice some additional information displayed on this visualization. Specifically a differentiation between household population and group quarters population. On exploration of the resources with the data set it seems that the group quarters population is population recorded living in environments such as dorms. We will loop back to an question that this data set introduced during the conclusion.
Properties
Next let's look at the properties by geographical regions. This is extracted from a Housing Stock Estimates data set also by the Metropolitan City Council. The housing stock data follows a similar annual structure as the population data however it is broken down year by year instead of a combined data set of Census data and Metropolitan Council Research data. The Properties data extends from 2005 to 2024 but does similarly breakdown by geographic region.
While reviewing this data you may wonder about the property type break down, visible by hovering over any point on the region timeline. For the most part the descriptions are self-evident however I want to note that Accessory Dwelling Units likely references attached homes, for example a self-contained living space on the same property as a main, single-family home. Such as backyard cottages, basement apartments, and in-law-suites.
Conclusion and Additional Questions
So why are we looking at this data? Well, I wanted to better understand the question of if there is a housing shortage or if there is another issue that may cause vacancies explored in earlier data stories. For the purpose of this analysis I will focus only on the population per household even though i would expect this to be not entirely reflective of population per group quarters. By taking the population of each geographic region and dividing it by the population per household we should, in theory, get an estimate of the number of truly occupied units.
One concern that has lingered at the back of my mind over this series of data analysis stories is that there are, for certain, more empty units than buildings actually classified as vacant. Units that are up to code but the landlord or owner has not filled or possibly second homes that remain vacant despite the constant struggle to find affordable housing.
Although simply dividing population by population per household averages and subtracting that from the actual sum of units is only an estimate it yields some interesting results.
As indicated in the visualization above I have estimated 12,827 units in Saint Paul may be vacant and 20,484 in Minneapolis, which raises the question. Why is rent so high if there are so many available places to live, that are in theory, up to code?
Another question that arose to the surface while I was evaluating population data is, how many people are living unhomed in our cities despite the unoccupied and vacant properties. I have made data requests to local shelters and to the Metropolitan City Council to see if they have an estimate of unhoused people in our cities to hopefully explore more aspects of this data.
If you have thoughts about this analysis or ideas on where I can find the missing data, feel free to leave a comment.
Once again, thank you for reading.
(11/18/2025) Addendum
I have since received a response from my inquiries on the representation of unhoused people in the population data sets. The data owners confirmed some Group housing population represents shelters and that some unhoused individuals still have a reported residency with family or friends. This mean a portion of the unhoused population should be accounted for in the group housing population and in the general population. The exact portion rate represented is unknown as its difficult to capture population numbers for individuals without a permanent residence. There were also efforts in 2020 to count the unhoused population however exactly where they were captured is uncertain. A likely explanation is population reported in group quarters for blocks where there are no apparent group quarters.
Feel free to let me know if you have additional thoughts or questions by utilizing the comments section below.
Comments