The analysis is focused on the two main typologies of buildings which are Multifamily Housing and Offices which accounts around 74 % of the buildings and 11.5 % respectively.
The LL84 was supplemented with PLUTO data which is a public dataset from Apple Bits to capture some physical and economic characteristics of the buildings. Therefore, it is formed by general characteristics, land use, and geographic data at the tax level. This dataset is building centric meaning that it contains features to describe both characteristics of the building and the area where located.
More specifically, PLUTO dataset is composed of 84 features from data by city agencies and contains 859,464 buildings distributed along the five boroughs. In Fig 2 , it can be seen the distribution of buildings contained in the data daset by boroughs.