Sunday, 17 March 2019

Digging deeper in the census data part 7: Building type and status

Hello and welcome to Open Citizen Data Science!

Today we will look into the last part of the census data, regarding building types (residential VS non residential), their composition and their status.
This last part is pretty important as it describes the physical landscape in many ways, allowing us to determine if we're looking at an old hamlet or part of an urban sprawl for example.

Let's take for example the center of the area with the most buildings:

Not what one could expect
It doesn't exactly look like an urban sprawl, but we're looking at an area with 1366 buildings in it.
The catch is that we're also looking at a very large area, which is a large part of a city territory:

Less densely populated areas tend to have bigger zones
Just like in many previous posts, this means that one metric is not enough to classify one area but we have to look at all possibile variables.
Speaking of which, let's do a recap of the available data in this section:

- Edifici e complessi di edifici - totale
- Edifici e complessi di edifici utilizzati
- Edifici ad uso residenziale
- Edifici e complessi di edifici (utilizzati) ad uso produttivo, commerciale, direzionale/terziario, - turistico/ricettivo, servizi, altro
- Edifici ad uso residenziale in muratura portante
- Edifici ad uso residenziale in calcestruzzo armato
- Edifici ad uso residenziale in altro materiale (acciaio, legno, ecc.)
- Edifici ad uso residenziale costruiti prima del 1919
- Edifici ad uso residenziale costruiti dal 1919 al 1945
- Edifici ad uso residenziale costruiti dal 1946 al 1960
- Edifici ad uso residenziale costruiti dal 1961 al 1970
- Edifici ad uso residenziale costruiti dal 1971 al 1980
- Edifici ad uso residenziale costruiti dal 1981 al 1990
- Edifici ad uso residenziale costruiti dal 1991 al 2000
- Edifici ad uso residenziale costruiti dal 2001 al 2005
- Edifici ad uso residenziale costruiti dopo il 2005
- Edifici ad uso residenziale con un piano
- Edifici ad uso residenziale con 2 piani
- Edifici ad uso residenziale con 3 piani
- Edifici ad uso residenziale con 4 piani o più
- Edifici ad uso residenziale con un interno
- Edifici ad uso residenziale con 2 interni
- Edifici ad uso residenziale da 3 a 4 interni
- Edifici ad uso residenziale da 5 a 8 interni
- Edifici ad uso residenziale da 9 a 15 interni
- Edifici ad uso residenziale con 16 interni o più
- Totale interni in edifici ad uso residenziale
- Edifici ad uso residenziale con stato di conservazione ottimo
- Edifici ad uso residenziale con stato di conservazione buono
- Edifici ad uso residenziale con stato di conservazione mediocre
- Edifici ad uso residenziale con stato di conservazione pessimo

This is enough to tell how many buildings, residential or not, of which material and age, their size and their status for each area.
As we're looking at building data, we will start by filtering out all areas without a building.

Let's take a look at the most populated area with the least number of buildings:

As usual, religious buildings will skew any attempt at finding lots of people in a small place
Just like many past attempts, we end with some sort of religious structure, so for this article we'll instead look at the number of buildings as driving metric. Let's move on and take a look at usage:

Another large rural area
With all 1345 buildings used and 1395 people in it, this is another case of perhaps questionable census area design, covering an area of over 50 square kilometers and several small towns. What is also questionable is that this is a 100% residential area according to the census.

Let's try to see the most non-residential area instead:

An industrial dock
This is better and it describes an industrial area pretty well. For once, the data seems also clean as thankfully there are no residents in it.

Let's now take a look at materials for residential areas and see what they mean.

Muratura portante:

This kind of construction type means mostly smaller buildings
Calcestruzzo armato:

From smaller houses to large complexes inspired by Le Corbusier, calcestruzzo is relatively cheap and fast
"Other" materials:

"Other" materials might be simply mis-reported but the top 10 seems nicer neighbourhoods
Looks like materials by themselves have too much variation to be useful alone.

Let's look at ages:

Before 1919:
This segment represents "historical" buildings.
1919-1945:

Mostly built between the wars, this segment represents buildings that survived WWII
1946-1960:

Built with various different styles, this group represents the post-war reconstruction effort
1961-1970:

Built in the "economic miracle" era, this group might be fancier in average
1971-1980:

With the highest growth ending, buildings started having a more "industrial" look, even fancier ones
1981-1990:

The trend continues towards the end of the cold war
1991-2000:

The end of the 20th century changes a little
2001-2005:

The new millennium seems to bring better ergonomics... or is there a bias in the data?
After 2005:

Smaller houses definitely got fancier with time
The "most buildings" selection criteria gives a bias towards zones with lots of small houses, so pictures are not very representative by themselves.

A distribution might be more useful:

Almost half of the residential housing has been built in 3 decades, with a marked slowdown after 1980, which is correlated with the slow-down of population growth.

What about number of floors?

2 floors seems to be the preferred option.

We also have the "interno" definition, let's see what it means visually:

1 interno:
Small houses
2 interni:
Bigger, scales faster than I expected
3-4 interni:
Definitely related to size
5-8 interni:
It might be related to stairwells
9-15 interni:
The higher the number, the wider the building
16+ interni:
Extremely wide, but not necessarily highly populated
The number of interni is pretty reliable in telling us how wide a building is (the largest categories are likely to span over multiple addresses) and in a way also telling us if the area is urban VS rural.

Let's see if the average number is helpful in giving us an estimate of density:

highest number:
This wouldn't be out of place in an Asian metropolis
lowest:

Pretty rural I'd say
Finally, let's see how they are distributed:

Looks like smaller houses are the most common in Italy after all!

What about their condition?

Ottimo (Excellent):
 
This applies to new and newly renovated buildings


Buono (Good):
In good condition, might need just a coat of paint
Mediocre:

The building is starting to show the signs of age, it will need to be renovated sooner rather than later
Pessime (Poor):
This definitely needs more than a coat of paint

Condition is a good indicator of social conditions, especially with older buildings. While many in the poor category have been renovated between 2011 and today, the fact that a building is allowed to reach such a poor state before fixing it means that the owners had to wait until they were pretty much forced to fix the problems rather than optimal maintenance.

This concludes our journey on the various building types. Stay tuned forour next article!

No comments:

Post a Comment