zenodoopen

Synthetic Indoor Climate and Occupancy Data from Office and Meeting Room Simulations

This is the dataset used for the publication "Coddora: CO2-based Occupancy Detection model trained via DOmain RAndomization". The goal is to provide training data for occupancy detection. The dataset contains one million days of data including 10 occupied days for each of 100,000 randomized room models (50,000 rooms considering office activity and 50,000 meeting room activity). Data were generated in EnergyPlus simulations according to the methodology described in the paper. When using the dataset, please cite: <blockquote> Manuel Weber, Farzan Banihashemi, Davor Stjelja, Peter Mandl, Ruben Mayer, and Hans-Arno Jacobsen. 2024. Coddora: CO2-Based Occupancy Detection Model Trained via Domain Randomization. In International Joint Conference on Neural Networks (IJCNN). June 30 - July 5, 2024, Yokohama, Japan. </blockquote> <h2>Dataset Structure</h2> The following files are provided:     1. dataset_office_rooms.h5   (provided as zip file)     2. dataset_meeting_rooms.h5   (provided as zip file)     3. simulated_occupancy_office_rooms.csv     4. simulated_occupancy_meeting_rooms.csv Please use an archiving tool such as 7zip to unzip the hdf5 files. Both hdf5 files contain two datasets with the following keys:     1. "data": contains the simulated indoor climate and occupancy data     2. "metadata": contains the metadata that were used for each simulation The csv files contain the time series of occupancy that were used for the simulations. <h2>Data</h2> Data includes the following fields: Datetime: day of the year (may be relevant due to seasonal differences) and time of the day Zone Air CO2 Concentration: CO2 level in ppm Zone Mean Air Temperature: temperature in °C Zone Air Relative Humidity: relative humidity in % Occupancy: level of occupancy relative to the maximum capacity of the room (in the range [0-1]) Ventilation: fraction of window opening in the range [0.01, 1] SimID: foreign key to reference the room properties the simulation was based on BinaryOccupancy: 0 or 1 denoting absence or presence (for binary classification)   Example row: <table> <tbody> <tr> <th>Datetime</th> <th>Zone Air CO2 Concentration</th> <th>Zone Mean Air Temperature</th> <th>Zone Air Relative Humidity</th> <th>Occupancy</th> <th>Ventilation</th> <th>simID</th> <th>BinaryOccupancy</th> </tr> <tr> <td> 10/09 11:21:00 </td> <td> 1084.5624647371608 </td> <td> 24.545635909907148 </td> <td> 41.18393114737054 </td> <td> 0.7 </td> <td> 0.0 </td> <td>99</td> <td>1</td> </tr> </tbody> </table> <pre> </pre> <h2>Metadata</h2> Metadata includes the following fields. Underscores denote that the field was not selected during randomization but calculated from the other values. width: room width in m length: room length in m height: hoom height in m infiltration:  infiltration per exterior area in m³/m²s outdoor_co2: co2 concentration in the outdoor air in ppm (set to a random value between [300, 500]) orientation: angle between the room's facade orientation and the north direction in degrees maxOccupants: room occupation limit, i.e. the maximum number of occupants _floorArea: floor area in m² (calculated from room dimensions) _volume: room volume in m³ (calculated from room dimensions) _exteriorSurfaceArea: surface area of the facade wall (calculated from room dimensions) _winToFloorRatio: ratio between total window area and floor area (calculated from room model) firstDayUsedOfOccupancySequence: selected starting day in the sequence of occupancy data for rooms with the respective maxOccupants value simID: unique identifier of the simulation to relate between simulation metadata and resulting simulated data   Example row: <table> <tbody> <tr> <th>width</th> <th>length</th> <th>height</th> <th>infiltration</th> <th>outdoor_co2</th> <th>orientation</th> <th>maxOccupants</th> <th>_floorArea</th> <th>_volume</th> <th>_exteriorSurfaceArea</th> <th>_winToFloorRatio</th> <th>firstDayOfUsedOccupancySequence</th> <th>simID</th> </tr> <tr> <td>5.481</td> <td>5.190</td> <td>3.264</td> <td>0.000214</td> <td>438.0</td> <td>316.0</td> <td>4.0</td> <td>28.446</td> <td>92.849</td> <td>16.940</td> <td>0.216</td> <td>192</td> <td>0</td> </tr> </tbody> </table>   <h2>Occupancy Data</h2> The occupancy data provided through the separate csv files contain the data from the upfront occupancy simulations that the climate simulation was based on. For each level of considered room occupancy limit (maxOccupants), the datasets provide minute values of occupancy throughout 1000 days. Datetime, Date, Timestamp: fictive time of simulated occupancy record (sequences are in 1-minute resolution) Occupants: number of present occupants Occupancy: binary occupancy state (0=unoccupied, 1=occupied) WindowState: binary state of ventilation (0=windows closed, 1=room is ventilated) maxOccupants: maximum number of occupants considered for the simulated sequence WindowOpeningFraction: fractional extent to which windows are opened, within the interval [0.01, 1] Example row: <table> <tbody> <tr> <th>Datetime</th> <th>Date</th> <th>Timestamp</th> <th>Occupants</th> <th>Occupancy</th> <th>WindowState</th> <th>maxOccupants</th> <th>WindowOpeningFraction</th> </tr> <tr> <td>2023-01-01 00:00:00</td> <td>2023-01-01</td> <td>1.672531e+09</td> <td>0</td> <td>0</td> <td>0</td> <td>1</td> <td>0.0</td> </tr> </tbody> </table>

ShareScore

40/100

Overall dataset sharing score

Score breakdown

These five areas show where the dataset supports — or may limit — practical reuse.

Stewardship: 8
Harmonization: 4
Access: 20
Reuse readiness: 8
Engagement: 0

Topics

CO2 Temperature Humidity EnergyPlus Simulation Buildings Occupancy Occupant Counts Indoor Climate