AllPriceMetadata.txt

Original columns: month, town, flat_type, block, street_name, storey_range, area_sqm, flat_model, lease_start, resale_price

month: Year and month in YYYY-MM format
year: Year in YYYY format
town: Town according to HDB planning areas (see under categorisation below for all of them). One town (Lim Chu Kang) only has transactions in the early 1990s while Punggol is a very new town without transactions in the early 1990s so you may encounter errors in some scenarios.
town_dummy: dummy variable from 1-6 based on my categorisation of towns from most to least prime. These are subjective, not objective so if you disagree with my categorisation and/or want to use your own then go ahead!
Categorisation:
north=['SEMBAWANG','YISHUN','WOODLANDS'] --> 5
northwest=['CHOA CHU KANG','BUKIT BATOK','BUKIT PANJANG','LIM CHU KANG'] --> 6
northeast=['HOUGANG','SENGKANG','PUNGGOL'] --> 4
primecentral=['BISHAN','BUKIT TIMAH', 'CENTRAL AREA'] --> 1
geocentral=['ANG MO KIO','TOA PAYOH','SERANGOON'] --> 2
westcentral=['CLEMENTI','QUEENSTOWN','BUKIT MERAH'] --> 2
eastcentral=['KALLANG/WHAMPOA', 'MARINE PARADE','GEYLANG'] --> 3
fareast=['BEDOK','PASIR RIS','TAMPINES'] --> 4
farwest=['JURONG EAST', 'JURONG WEST'] --> 5
flat_type: HDB-specified flat type ['4 ROOM' '3 ROOM' '1 ROOM' '5 ROOM' 'EXECUTIVE' '2 ROOM'
'MULTI GENERATION']
block: Block number (NOT flat number) of the HDB block
street_name: Street name of the HDB block
address: Identifying address formed by concatenating block+street_name+"SINGAPORE". No postal code unfortunately but I may try to add it in the future.
latitude: Latitude coordinate of the HDB block, mostly between 1.27-1.46
longitude: Longitude coordinate of the HDB block, mostly between 103.5-104.0
storey_range: 3-number ranges (e.g. 1 TO 3, 19 TO 21) that contain the actual storey number (HDB doesn't reveal the exact floor)
storey: Middle number in the storey range, useful as a numerical approximation that can be used for data visualisation/machine learning
area_sqm: Size of the flat in square metres
flat_model: HDB-specified flat model (not the same as flat type) ['New Generation', 'Improved', 'Standard', 'Model A', 'Apartment', 'Maisonette', 'Model A-Maisonette', 'Simplified', 'Terrace', 'Improved-Maisonette', 'MULTI GENERATION', 'Premium Apartment', 'Multi Generation', 'Adjoined flat', 'Premium Maisonette', '2-room', 'Model A2', 'DBSS', 'Type S1', 'Type S2', 'Premium Apartment Loft']
lease_start: Year in which the 99-year lease commenced.
lease_rem: The number of years of remaining lease, as calcualted from 2021-lease_start. For a few anomalies, the year of transaction is BEFORE the year of lease starting which is strange but could be due to house flipping before the Minimum Occupancy Period was introduced?
resale_price: The agreed transaction price in nominal S$ of the house, excluding stamp duties and other agent fees
price_psm: The price per square metre as calculated from resale_price/area_sqm
price_psm_yearly: The "yearly" price per square metre as calculated from price_psm/lease_rem --> removed in Version 2
Core CPI: The core CPI index value at the month of the transaction, based on the data by MAS
price cpi_adj: The real, inflation-adjusted resale price of the flat, calculated using (resale_price/Core CPI)*100
price_psm cpi_adj: The real, inflation-adjusted price per square metre of the flat, which is the best indicator of how expensive the flat is. Calculated using (price_psm/Core CPI)*100
bala lease pct: The HDB's land value percentage of freehold land value, referencing Bala's Table. Important to consider because when the 99-year lease expires, the HDBs technically have 0 value. Adjusting for the lease remaining in the same way as inflation-adjusting, you can consider the price that is implied if the flat were to have a full 99-year lease.
price lease_adj implied: The resale price implied if the flat were to have a full 99-year lease, calculated from (resale_price/bala lease pct)*96
price_psm lease_adj implied: The price per square metre implied if the flat were to have a full 99-year lease, calculated from (price_psm/bala lease pct)*96
price cpi_lease_adj implied: The inflation-adjusted resale price implied if the flat were to have a full 99-year lease, calculated from (price cpi_adj/bala lease pct)*96
price_psm cpi_lease_adj implied: The inflation-adjusted price per square metre implied if the flat were to have a full 99-year lease, calculated from (price_psm cpi_adj/bala lease pct)*96
year_gni: Singapore's GNI per capita (nominal S$) in the year of the transaction