OBPlatform
A package to interact and download behavior data from ASHRAE Global Occupant Behavior Database. Currently available on PyPI and conda-forge. More features coming in the furture.
Features
List all behavior types available in the database.
Download data archive (ZIP file) based on behavior type and study id inputs (with progress bar).
Query studies based on (behaviors, countries, cities, (building type + room type))
Query available behavior types based on study ids
Installation
poetry
poetry install
pip
pip install --upgrade obplatform
conda
conda install -c conda-forge obplatform
Note
For Python 3.10: If you see an error like the following when resolving dependencies, it’s caused by a bug in conda with Python 3.10.
Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
PackagesNotFoundError: The following packages are not available from current channels:
- python=3.1
Three possible solutions:
Create a new conda environment with Python <3.10.
Upgrade conda to a new version. (conda released 4.11.0 on 11/22/2021 at GitHub, which fixed the bug for Python 3.10. However, it will still take some time before conda 4.11.0 is available on Anaconda Cloud).
Use mamba, which is a reimplementation of the conda package manager in C++. It is much faster and contains less bugs.
mamba
Once you activate the environment through conda or micromamba:
mamba install -c conda-forge obplatform
Example
import logging
import zipfile
import pandas as pd
from obplatform import Connector, logger
connector = Connector()
# List all behaviors available in the database
print(connector.list_behaviors())
# Print progress information
# Comment out the following line to hide progress information
logger.setLevel(logging.INFO)
# Download Plug Load + Occupant Presence behaviors from study 22, 11, and 2.
connector.download_export(
"data.zip",
["Plug_Load", "Occupancy_Measurement"],
["22", "11", "2"],
show_progress_bar=True, # False to disable progrees bar
)
behavior_type = "Plug_Load"
study_id = "22"
zf = zipfile.ZipFile("data.zip")
df = pd.read_csv(zf.open(f"{behavior_type}_Study{study_id}.csv"))
print(df.head())
# List all behaviors available in study 1, 2, 3, and 4
json_study_behaviors = connector.list_behaviors_in_studies(studies=["1", "2", "3", "4"])
print(json_study_behaviors)
# List all studies available in the database, filtered by behavior types,
# countries, cities, {building type, room_type} combinations.
json_studies = connector.list_studies(
behaviors=["Occupancy_Measurement", "Plug_Load"],
countries=["USA", "UK"],
cities=["Palo Alto", "Coventry", "San Antonio"],
buildings=[
{
"building_type": "Educational",
"room_type": "Classroom",
},
{
"building_type": "Educational",
"room_type": "Office",
},
{
"building_type": "Residential",
"room_type": "Single-Family House",
},
],
)
print(json_studies)
Usage
Available behavior types
Please only use the following names as input. e.g. Please use
Lighting_Status
(listed below) instead of
Lighting Adjustment
(displayed on the website).
'Plug_Load', 'Fan_Status', 'Door_Status', 'HVAC_Measurement', 'Lighting_Status', 'Occupant_Number', 'Occupancy_Measurement', 'Other_HeatWave', 'Other_Role of habits in consumption', 'Other_IAQ in Affordable Housing', 'Shading_Status', 'Window_Status'
In the next version, the package will auto detect either type of input and convert to the correct query parameter.
Note
Study 2 is a special case. It has very large source files (> 2 GB) so we
compressed all data in study 2 as a single .tar.gz
file. In the
example above, data.zip
contains a tar.gz
file along with
several separate csv files from other studies. When writing libraries to
read from csv file from the downloaded zip, Study 2 should be treated as
a special case.
Changelog
2021-11-18: Release 0.1.3
2021-11-19: Release 0.1.4, fixed a minor issue with Python 3.10.0
2021-11-23: Release 1.0.0
Breaking changes:
Behavior type (query field) “Occupancy” has been renamed to “Occupancy_Measurement” to keep the name consistent. The example above has been changed accordingly. The server will reject query field “Occupancy”.
Added endpoint to check backend server health
Added endpoint to query available behavior types based on Study IDs
2021-12-01: Release 1.1.0
Added endpoint to query available studies based on (behaviors, countries, cities, (building type + room type))
2021-12-10:
Breaking change: renamed Appliance Usage to Plug Load on the server end. The example code has been changed accordingly.