For example, I have the percentage breakdown of various building typologies and with them the estimated price of these buildings in £/m2 (see table below). So my question is whether it is possible to estimate the £/m2 for a building based only on the typologies.
Some potential methods that I thought could work are as follows:
Set up a system of simultaneous equations. However, this would be difficult when considering hundreds of buildings and tens of building typologies.
Determine an average £/m2 for a % of each typology type (e.g. 1% of retail space is ~£2/m2).
Find the best matching building. It is clear that Building 1 has the closest typology to Building 5. However, this can be tedious to do for hundreds of buildings.
A mixture of the above, i.e. grouping similar buildings together than deriving an average £/m2 per group.
Is there a more precise soultion to approach this promblem whilst consdering the constraints of the large dataset and data variabiltiy?
| Retail | Residential | Industrial | Car park | Green open space | £/m2 | ||
|---|---|---|---|---|---|---|---|
| Building 1 | 15% | 55% | 0% | 10% | 20% | 100 | |
| Building 2 | 10% | 0% | 70% | 10% | 10% | 80 | |
| Building 3 | 0% | 70% | 0 | 10% | 20% | 90 | |
| Building 4 | 70% | 0% | 0% | 20% | 10% | 110 | |
| Building 5 | 10% | 60% | 0% | 5% | 25% | ? |