ASHRAE Global Thermal Comfort Database

The ASHRAE Global Thermal Comfort Database II combines sets of objective indoor environmental measurements with accompanying "right-here-right-now" subjective evaluations by occupants from buildings around the world. The database is intended to support diverse inquiries about thermal comfort in field settings. This is the official repository which contains the code used to maintain the Database II from version 2.1 onward. You can download the latest version of the database from the Github repository or Dryad.

This repository contains the source code we used to generate each new version of the dataset. This allows you to track the changes we have implemented. We encourage you to submit an issue if you spot any errors or you would like us to implement some changes. The repository comprises different folders, each containing the source code and data we used to generate a new version of the database.

Columns names and description

Database Metadata

parameter	description
building_id	Unique building identifier [integer]
building_id_inf	Flag indicating if unique building identifier was from original data source [no] or inferred [yes] from unique groupings of publication, country, city, season, building type, and cooling type
contributor	Principal contact person regarding the data
publication	Published paper describing the project from where the data was collected
region	Region of field study
country	Country of field study
city	City of field study
lat	Latitude of city [°]
lon	Latitude of city [°]
climate	Type of climate according to Köppen climate classification
building_type	Type of building [office, multifamily housing, classroom, senior center, other]
cooling_type	Cooling strategy of building [air conditioned, mixed mode, naturally ventilated]
year	Year of field study [yyyy]
records	Number of records for that building ID
has_age	Flag indicating if there age was recorded [yes, no] and if it was a categorical variable in the original data source [categorical]
has_ec	Flag indicating if environmental controls were in the original data source [yes, no]
has_timestamp	Flag indicating if measurement timestamp was in the original data source [yes, no]
timezone	IANA time zone of field study
met_source	Source of meteorological data for t_out and rh_out [ghcn_d = from GHCN-D, original_data = from original data source; rp884 = from RP884 database]
isd_station	ISD station code for t_out_isd, rh_out_isd and t_mot_isd
isd_distance	Estimated distance of ISD station to city of field study [km]
database	Version of database when data source was added [1, 2, 2.1]
quality_assurance	Flag indicating if dataset from contributor passed automated quality assurance check [pass, fail]

Database Measurements

parameter	description
building_id	Unique building identifier [integer]. Note: some building IDs are inferred - see building_id_inf in metadata table
timestamp	Timestamp of measurement [yyyy-mm-dd]
season	Season measurement was made [summer, winter, hot/wet, cool/dry]. Note: based on the following assumptions when timestamp and location are known: northern hemisphere latitudes <20 are hot/wet from May-Oct and cool/dry from Nov-Apr; northern hemisphere latitudes >=20 are summer May-Oct and winter from Nov-Apr; vice versa for Southern Hemisphere
subject_id	Unique subject identifier for future studies with repeat samples [integer]
age	Age of subject [years]. Note: some studies used age ranges instead of years - see has_age in metadata table
gender	Gender of subject [female, male]
ht	Height of subject [m]
wt	Weight of subject [kg]
ta	Air temperature measured in the occupied zone [°C]
ta_h	Air temperature measured at 1.1 m above the floor [°C]
ta_m	Air temperature measured at 0.6 m above the floor [°C]
ta_l	Air temperature measured at 0.1 m above the floor [°C]
top	Operative temperature calculated for the occupied zone [°C]
tr	Radiant temperature measured in the occupied zone [°C]
tg	Globe temperature measured in the occupied zone [°C]
tg_h	Globe temperature measured at 1.1 m above the floor [°C]
tg_m	Globe temperature measured at 0.6 m above the floor [°C]
tg_l	Globe temperature measured at 0.1 m above the floor [°C]
rh	Relative humidity [%]
vel	Air speed measured in the occupied zone [m/s]
vel_h	Air speed measured at 1.1 m above the floor [m/s]
vel_m	Air speed measured at 0.6 m above the floor [m/s]
vel_l	Air speed measured at 0.1 m above the floor [m/s]
vel_r	Relative air speed used to calculate the PMV [m/s]
met	Average metabolic rate of the subject [met]
clo	Intrinsic clothing ensemble insulation of the subject [clo]
clo_d	Dynamic clothing, used to calculate the PMV [clo]
activity_10	Average metabolic rate of the subject in the last 10 minutes [met]
activity_20	Average metabolic rate of the subject in the last 20 minutes [met]
activity_30	Average metabolic rate of the subject in the last 30 minutes [met]
activity_60	Average metabolic rate of the subject in the last 60 minutes [met]
thermal_sensation	Vote on the ASHRAE thermal sensation scale [-3 (cold) to 0 (neutral) +3 (hot)]
pmv	Predicted mean vote, calculated in compliance with the ISO 7730
pmv_ce	Predicted mean vote, calculated in compliance with the ASHRAE 55 2020
ppd	Predicted percentage dissatisfied [%] calculated in compliance with the ISO 7730
ppd_ce	Predicted percentage dissatisfied [%] calculated in compliance with the ASHRAE 55 2020
set	Standard effective temperature [°C]
thermal_acceptability	Thermal acceptability [acceptable, unacceptable]
thermal_preference	Thermal preference [cooler, no change, warmer]
thermal_comfort	Thermal comfort [1 (very uncomfortable) to 6 (very comfortable)]
air_movement_acceptability	Air movement acceptability [acceptable, unacceptable]
air_movement_preference	Air movement preference [less, no change, more]
blind_curtain	State of blinds or curtains [0 = open; 1 = closed]
fan	State of fan [0 = off, 1 = on]
window	State of window [0 = open, 1 = closed]
door	State of doors [0 = open, 1 = closed]
heater	State of heater [0 = off, 1 = on]
t_out	Outdoor air temperature from original dataset [°C]
rh_out	Outdoor relative humidity from original dataset [%]
t_out_isd	Average daily outdoor air temperature from ISD [°C]
rh_out_isd	Average relative humidity from ISD [%]
t_mot_isd	Calculated 7-day running mean outdoor temperature [°C]

Contributing

You can contribute to the project by sending us your dataset. Please clean and collate the data into a single .csv file that matches the existing DBII structure outlined. This simple web tool will help error check and format the data. It also has a downloadable .csv template file. Note that this tool reflects the old database format - we will work to update the tool soon.

Reporting issues

You can report issues with the database by using this link

Changelog

Version 2.1

This is the first major update to the ASHRAE Global Thermal Comfort Database II since it was published in 2018. In the time since release, it has been used by hundreds of researchers who have found errors and requested new features. This update aims to address some of those while improving operability with other resources. As a result, there are breaking changes.

The major updates to the database are:

1. 6,566 new records from field studies in India, Malaysia, Italy, Japan, Indonesia, Singapore, and Cyprus. Thanks to Hasim Altan (Arkin University of Creative Arts and Design), Siti Aisyah Damiati (University of Adelaide), Harimi Djamila (Universiti Malaysia Sabah), Jeetika Malik (Lawrence Berkeley National Laboratory), Bertug Ozarisoy (University of East London), Elisabetta Maria Patane (University of Bath), and Rajan Rawal (CEPT University) for kindly sharing their data with the community.
2. Improvements to the meteorological data. In the earlier version of the database, the source and quality of the meteorological data was unclear in some cases. We now use the NOAA Integrated Surface Database (ISD) for meteorological data where possible. Thanks to Peixian Li (Tongji University) for adding the timestamps from the original datasets (where available). Meteorological data from either the original dataset or the ISD are now included along with the relevant metadata.
3. The database is now split into a metadata table and a measurements table. This should make working with the full dataset quicker and easier due to smaller file sizes and new fields in the metadata table.

There are many other changes included in this update. The full change log is below:

• Added building ID from RP-884 dataset
• Added timestamp from RP-884 dataset
• Added meteorological data from RP-884 dataset
• Added timestamp to datasets where known [Ariel Li - Tongji University]
• Added 6,566 measurements from new data contributions
• Split database into metadata and measurement tables
• Recalculated PMV/PPD using pythermalcomfort
• Performed simple quality assurance on measurement database
• Shared python code used to calculate PMV/PPD
• Shared R code with an example analysis of adaptive thermal comfort
• Dropped records with missing ta
• Recoded many categorical scale variables to standardised English label text
• Dropped all imperial units
• Renamed headers to remove spaces and special characters
• Renamed headers for PMV inputs to align with software packages [Federico Tartarini - BEARS; Marcel Schweiker - RWTH Aachen University]
• Recoded cooling types Mechanically ventilated; and Mixed (hybrid) into mixed mode
• Recoded season when timestamp and location is known
• Added average daily temperature calculated from ISD data
• Added running mean outdoor temperature from ISD data
• Added inferred building id based on unique groupings of publication, country, city, season, building type, and cooling type
• Added inferred building ID flag to metadata table
• Added sample size for each building ID to metadata table
• Added approximate coordinates to metadata table
• Added time zone to metadata table
• Added source of meteorological data to metadata table
• Added ISD station number and distance to metadata table
• Added categorical age field flag to metadata table [Jing Xiong - The University of Sydney]
• Added environmental control flag to metadata table
• Added timestamp flag to metadata table
• Added region to metadata table
• Added quality assurance flag to metadata table
• Replaced full reference in metadata table with DOI where known
• Standardised country names in metadata table
• Fixed season in Oseland dataset [Carlucci; Matteo Favero - NTNU]
• Added building ID in Oseland dataset
• Fixed asv and building type in Zhangeri dataset [Salvatore Carlucci - The Cyprus Institute; Matteo Favero - Norwegian University of Science and Technology]
• Added building ID in Zhangeri dataset
• Removed thermal comfort vote for Zhangeri dataset [Salvatore Carlucci - The Cyprus Institute; Matteo Favero - Norwegian University of Science and Technology]
• Added unique identifier for records (record_id) to help error reporting
• Added new column (subject_id) for future submissions with repeated measures