Bytero Rent Grid
The monthly rent of a flat from our rent model, computed in advance for every 1 km square, by rooms, floor area and condition.
What a row is
One row is one square and one flat: the square (CELL_ID with its centre), the flat (ROOMS, AREA_M2, CONDITION), and the modelled rent per month without utilities (RENT_NET_EUR) and with utilities (RENT_GROSS_EUR), with a range around the rent without utilities.
How the values are made
The rent model learns from the asking rents of rental adverts. Adverts state a rent in two ways, with or without utilities; the model is trained on the rent without utilities and adds the utilities for the second figure.
For each square the model is run at one point where flats are offered, for five flat sizes and three conditions. Squares with fewer than 10 rental adverts in the last 365 days are left out.
The squares carry the same ids as the Bytero Engine value grid, so the rent and the value of a flat join on CELL_ID.
The range
The range is measured. Adverts whose text states the rent without utilities are split into two halves: one half gives the band that holds 80 % of the differences between an advert and its row of the grid, and RANGE_COVER_PCT is the share of the other half that fell inside that band.
What it is not
The values are modelled asking rents, not agreed rents. A row is one point per square and a rounded flat, so a single flat can differ from its row. Flats in original condition are not in the grid: too few are offered for rent for the model to price them.
Table FLATS.RENT_VALUES
18 columns
Bytero rent model: monthly rent of a flat, run for every 1 km square over rooms, floor area and condition. Modelled asking rents.
COUNTRYtext- ISO 3166 code
REGIONtext- Region the grid covers
DISTRICTtext- District most adverts in the square name
CELL_IDtext- The 1 km square; the same ids as the Bytero Engine value grid
CELL_LATnumber- Latitude of the square centre (WGS84)
CELL_LONnumber- Longitude of the square centre (WGS84)
ROOMSnumber- Rooms of the flat valued
AREA_M2number- Floor area of the flat valued, m2
CONDITIONtext- Condition of the flat valued: New build, Fully renovated, Partly renovated
RENT_NET_EURnumber- Modelled asking rent per month in EUR, without utilities
RENT_GROSS_EURnumber- Modelled asking rent per month in EUR, with utilities
RENT_NET_EUR_M2number- RENT_NET_EUR per m2
RENT_NET_LOW_EURnumber- Low end of the range of the rent without utilities
RENT_NET_HIGH_EURnumber- High end of that range
RANGE_COVER_PCTnumber- Share of test adverts whose rent without utilities fell inside the range, read through the grid
ADVERTS_365Dnumber- Flat rental adverts seen in the square in the last 365 days
MONTHdate- Month the values were computed
MODELtext- Model that computed the value






