PreprintPDF Available

Abstract and Figures

Internet-of-Things (IoT) devices in buildings and wearable technologies for occupants are quickly becoming widespread. These technologies provide copious amounts of high-quality temporal data pertaining to indoor and outdoor environmental quality, comfort, and energy consumption. However, a barrier to their use in many applications is the lack of spatial context in the built environment. Adding Building Information Models (BIM) and Geographic Information Systems (GIS) to these temporal sources unleashes potential. We call this data convergence the Internet-of-Buildings or IoB. In this paper, a digital twin case study of data intersection from various systems is outlined. Initial insights are discussed for an experiment with 17 participants that focused on the collection of occupant subjective feedback to characterize indoor comfort. The results illustrate the ability to capture data from wearables in the context of a BIM data environment.
Content may be subject to copyright.
Preprint accepted at CISBAT 2021 - Carbon Neutral Cities - Energy Efficiency & Renewables
in the Digital Era, EPFL, Lausanne, Switzerland, 8-10 September 2021
The Internet-of-Buildings (IoB) – Digital twin
convergence of wearable and IoT data with GIS/BIM
Clayton Miller1,, Mahmoud Abdelrahman1, Adrian Chong1, Filip
Biljecki2,3, Matias Quintana1, Mario Frei1, Michael Chew1, Daniel
1Dept. of the Built Environment, National University of Singapore (NUS), Singapore
2Dept. of Architecture, National University of Singapore (NUS), Singapore
3Dept. of Real Estate, National University of Singapore (NUS), Singapore
Abstract. Internet-of-Things (IoT) devices in buildings and wearable technologies for
occupants are quickly becoming widespread. These technologies provide copious amounts of
high-quality temporal data pertaining to indoor and outdoor environmental quality, comfort,
and energy consumption. However, a barrier to their use in many applications is the lack of
spatial context in the built environment. Adding Building Information Models (BIM) and
Geographic Information Systems (GIS) to these temporal sources unleashes potential. We
call this data convergence the Internet-of-Buildings or IoB. In this paper, a digital twin case
study of data intersection from various systems is outlined. Initial insights are discussed for an
experiment with 17 participants that focused on the collection of occupant subjective feedback
to characterize indoor comfort. The results illustrate the ability to capture data from wearables
in the context of a BIM data environment.
1. Introduction
The Internet-of-Things (IoT) paradigm has driven the deployment of a vast array of sensors
and devices that are designed to improve the performance of buildings. There are several key
challenges that IoT networks are meant to address. For example, there is a 50:50 chance of an
occupant being comfortable in a building according to a survey of 52,980 occupants in 351 office
buildings, which found that 50% were dissatisfied with their indoor environment [1]. From the
operations and maintenance side, likely, 16% of the energy used in the building could easily be
saved using no and low-cost optimization and commissioning [2]. In addition to IoT, professionals
in the built environment, such as architects, engineers, and urban planners, have also been
digitizing spatial data at a rapid pace through the use of 3D virtual information tools such
as building and city information and energy modeling (GIS/BIM/BEM) [3], and a significant
amount of effort is going into determining uses in the operations phase [4].
This paper focuses on the creation of a data aggregation platform to test how Internet-of-
Things (IoT) and wearable technologies can be better converged with Building Information
Models (BIM) and Geographic Information Systems (GIS). Developing and training Artificial
Intelligence (AI) models to make predictions based on these data can have an impact on
occupancy detection, energy systems optimization, and human comfort and wellness [5, 6, 7].
This process is done through the development of the Internet-of-Buildings (IoB) data convergence
platform on the campus of the National University of Singapore (NUS). These technologies
demonstrate the value of synthesizing data from numerous sources and creating interfaces that
enable building operations professionals to make decisions. This work attempts to meet the
deficiencies in previous literature [8] through the synthesis of data from dynamic IoT sources
such as wearables into GIS and BIM environments.
2. The Internet-of-Buildings (IoB)
The focus of this research is the development and testing of a data convergence platform. Figure
1 illustrates this new paradigm and the potential capabilities within various use case scenarios.
First, data connectivity is enhanced with the introduction of 5G capabilities indoors and outdoor,
increasing the bandwidth and reducing data exchange latency. Next, the maintainability of
buildings is improved due to the use of AI-driven drones that can scan buildings for maintenance
issues using the BIM/GIS models as a context. In universities, using the building as a teaching
test-bed is enhanced using spatial and temporal data convergence through wearable and mobile
augmented and virtual reality devices (AR/VR). Energy systems and digital twin building energy
models are calibrated using high-resolution occupancy data, resulting in the optimization of
climate control and lighting systems [9]. Indoor and outdoor wellness and thermal comfort
is enhanced using data collected from multiple sources and used to inform and even modify
energy consumption behavior to achieve conservation targets. Energy behavior is collected by
synthesizing occupant information with smart meter data and social and behavioral intervention
methods that are tested in a field setting. Wellness, productivity, and space utilization are
evaluated at the zone level, and people can be given recommendations about the zones that
would result in the most comfortable and productive experience for them. A key feature for tying
together the IoT/wearable and spatial domain is the ability to locate where a person or device
Data Connectivity 4G/5G/Wifi for Smart
Safety, and Project
5G, Drones, and BIM for Maintainability,
Decision-support, and Safety analysis
Teaching and
Smart Devices,Wearable IoT and AR/VR for
education of students
Energy Efficiency,
Conservation, and
BMS/Smart Meter, PV Meters and BIM/BEM
integration to optimize control and find system
Indoor and Outdoor
Wellness and
Thermal Comfort
Smart Devices, Indoor and Outdoor
IoT and Wearable IoT for Subjective
Indoor and Outdoor
Localization and
Occupant Behavior
GPS enhanced with GIS, BIM, and WiFi and
Bluetooth localization and occupant detection
Privacy and
IoT Contextual Pairing technologies
supplement security and privacy
The Internet-of-Buildings (IoB) Framework
Wifi Router
Global Positioning
System (GPS)
4G and 5G
Smart Devices
Air, Sound,
Light, and
IoT Sensors
Building Information
Model (BIM)
and Digital Twin Building
Energy Model (BEM)
Information Systems
Smart Meter
PV Smart
Figure 1: The deployment of IoT in the built environment combined with spatial and human
data sources creates the IoB Framework.
in the building context beyond the accuracy that GPS can achieve. This situation requires the
use of indoor localization technologies such as those powered by Bluetooth. Security is enhanced
through contextual pairing with other sensors that add more layers to the authentication
process for digital access control systems. Finally, privacy is improved as the access to personal
information for IoB is not controlled in a centralized way; smartphones and wearable IoT owners
are able to choose what information is shared through the application interface.
3. Case study deployment
The case study implementation in this paper focuses on the aggregation and synthesis of data
from sensors located at various parts of the NUS campus, including the SDE4 building, a 5G-
enabled, net-zero energy building. Figure 2 illustrates the various data sources that have been
converged or are in progress of being connected. The first pillar focuses on the collection of
data from humans through wearable devices, an example of which is when smartwatches were
deployed to collect physiological and subjective feedback data from building occupants [10].
The following two pillars fall within the category of 5G-connected devices such as smartphones,
robots, and drones [11]. The next pillar focuses on collecting spatial information from occupants
and objects using an indoor localization system based on Bluetooth beacons deployed across the
test case buildings. This spatial information is then converged with a digital twin of the building
extracted from the BIM. The last two pillars focus on the data convergence of fixed smart sensors
from indoor (building-scale) and outdoor (urban-scale) environments. These are sensors, for
IoB Data Convergence with BIM/BEM/GIS using a Spatially-aware Time-Series Database
Wearable Data Dynamic 5G Connected
Fixed IoT Sensor
Ingestion Server
BMS Server
Virtual Campus
Ingestion Server
IoT Server
Dynamic IoT
Lambda Server
Drone Hardware
Lambda Server
Mobile Data
Ingestion Server
(Outdoor) IoT
(Indoor) IoT
Figure 2: The development of the IoB platform focuses on the data convergence of data from
various projects and systems in terms of data sources (columns) and stages of data processing
and convergence (rows).
example, that measure environmental air quality and thermal conditions across campus. The
merging of these data sets is possible through a series of technologies from third parties such
as hardware devices, middleware servers, and event-driven lambda function services. Finally,
the data convergence occurs in a time-series database that indexes the various sources using the
spatial context.
3.1. Dynamic wearable data convergence with BIM
Figure 3 shows this methodology from an experiment that used the IoB to collect and converge
data related to comfort preferences [12]. It shows the path of a human across the floor plan of
the SDE4 building at NUS (left). As the occupant traverses the space, IoB collects physiological
information such as heart rate and step count while using indoor localization to calculate the
proximity to various objects in the building. These dynamic relationships are calculated using
BIM convergence with the IoT and human-generated data. The illustration shows a fish-eye
view of the occupant (bottom center) and the indoor localization app being used to track the
path. Finally, the visualization shows a vector-based proximity diagram (right) that illustrates
the relationship of the human and the various spatial objects in a vector model [13].
3.2. Thermal comfort preference experiment
For six weeks in early 2021, an experiment was deployed in the SDE1, SDE2, and SDE4 buildings
with the use of the Cozie Fitbit smartwatch platform ( to
collect subjective thermal preference feedback. This method builds upon previous work [12, 14],
but with adaptations to increase the number of ecological momentary assessment questions and
the diversity of spaces in the case study buildings. The test participants were asked about their
thermal preference, clothing level, and activity level on the Cozie watch-face several times per
day while they were in the case study buildings. Their responses were attached to the BIM
through a Bluetooth localization app that was also installed on their phone.
Figure 3: An example of the data convergence from BIM, IoT, BEM, and human-generated
data in the Internet-of-Buildings platform. A full animated demo is found at this link:
4. Results and discussion
Figure 4 shows the results of the experimental implementation. The top row of the figure
illustrates the data collected in the spatial context in three levels of the case study building.
The IoB platform facilitated the capture of physiological and subjective feedback data from a
smartwatch and connected it to the spatial context of the BIM environment. The charts in
the next two rows show an overview of the preliminary analysis of this data set towards the
characterization of thermal preference in these buildings.
The use of the IoB platform to merge the subjective feedback data with the spatial context
resulted in several key insights. Many of the zones were outdoor spaces, and the results show a
higher than usual number of prefer cooler feedback, especially from the male participants. The
point clouds developed in the various regions of the floor plan illustrate the potential hot spots
in which there could be causes of discomfort not captured by the temperature and humidity
sensors installed. These initial results set the foundation for further analysis in contextualizing
the feedback according to each point’s relationship with the objects around it. There was an
expected imbalance in the feedback, which can be investigated using techniques to enhance the
data set before prediction [15]. There is further work planned that falls outside the scope of this
Figure 4: Data collected during experiments related to characterize different parts of the spatial
context for thermal preference for three levels in the case study buildings (SDE1-4). The floor
plans of each level illustrate the locations of subjective data collection (top) while the histograms
(middle and bottom) show a breakdown of occupant feedback according to gender and clothing
levels. This figure is best viewed in color.
paper to use these data for modeling techniques that can predict which zones would best fit the
preferences of individual occupants.
5. Conclusion
This paper outlines the creation and testing of a data convergence platform for IoT/wearable
and BIM/GIS data. This methodology converges data for this case study from indoor and
outdoor sensor networks and occupants using smartwatch and smartphone interfaces. These
data are combined with the BIM of the case study using a mapping process between the spatial
and temporal contexts. Data collected from a deployment with 17 experimental participants
focused on the collection of thermal comfort preference subjective feedback. The data from
this experiment were converged on the IoB platform in the BIM context and showed insight
into thermal comfort preference in different spatial zones. The next phase is to train machine
learning models for occupant preference prediction in the spatial context provided by the BIM.
[1] M. Frontczak, S. Schiavon, J. Goins, E. Arens, H. Zhang, P. Wargocki, Quantitative relationships between
occupant satisfaction and satisfaction aspects of indoor environmental quality and building design, Indoor
Air 22 (2) (2012) 119–131.
[2] E. Mills, Building commissioning: a golden opportunity for reducing energy costs and greenhouse gas
emissions in the united states, Energ. Effic. 4 (2) (2011) 145–173.
[3] F. Biljecki, J. Lim, J. Crawford, D. Moraru, H. Tauscher, A. Konde, K. Adouane, S. Lawrence, P. Janssen,
R. Stouffs, Extending CityGML for IFC-sourced 3D city models, Autom. Constr. 121 (2021) 103440.
[4] J. Lim, P. Janssen, F. Biljecki, Visualising detailed CityGML and ade at the building scale, ISPRS - Int.
Arch. Photogramm. Remote Sens. Spat. Inf. Sci. XLIV-4/W1-2020 (2020) 83–90.
[5] M. M. Abdelrahman, S. Zhan, C. Miller, A. Chong, Data science for building energy efficiency: A
comprehensive text-mining driven review of scientific literature, Energy Build. 242 (110885) (2021) 110885.
[6] S. Altomonte, J. Allen, P. M. Bluyssen, G. Brager, L. Heschong, A. Loder, S. Schiavon, J. A. Veitch, L. Wang,
P. Wargocki, Ten questions concerning well-being in the built environment, Build. Environ. 180 (2020)
[7] F. Stazi, F. Naspi, M. D’Orazio, A literature review on driving factors and contextual events influencing
occupants’ behaviours in buildings, Build. Environ. 118 (2017) 40–66.
[8] B. Dave, A. Buda, A. Nurminen, K. Fr¨amling, A framework for integrating BIM and IoT through open
standards, Autom. Constr. 95 (2018) 35–45.
[9] A. Chong, G. Augenbroe, D. Yan, Occupancy data at different spatial resolutions: Building energy
performance and model calibration, Appl. Energy 286 (2021) 116492.
[10] P. Jayathissa, M. Quintana, T. Sood, N. Nazarian, C. Miller, Is your clock-face cozie? a smartwatch
methodology for the in-situ collection of occupant comfort data, J. Phys. Conf. Ser. 1343 (1) (2019)
[11] M. Y. L. Chew, E. A. L. Teo, K. W. Shah, V. Kumar, G. F. Hussein, Evaluating the roadmap of 5G
technology implementation for smart building and facilities management in singapore, Sustain. Sci. Pract.
Policy 12 (24) (2020) 10259.
[12] P. Jayathissa, M. Quintana, M. Abdelrahman, C. Miller, Humans-as-a-Sensor for Buildings—Intensive
longitudinal indoor comfort models, Buildings 10 (10) (2020) 174.
[13] M. Abdelrahman, A. Chong, C. Miller, Build2Vec: Building representation in vector space, in: Symposium
on Simulation in Architecture + Urban Design, SimAUD 2020, 2020, pp. 101–104.
[14] P. Sae-Zhang, M. Quintana, C. Miller, Differences in thermal comfort state transitional time among comfort
preference groups, in: 16th Conference of the International Society of Indoor Air Quality and Climate:
Creative and Smart Solutions for Better Built Environments, Indoor Air 2020, 2020, p. 166587.
[15] M. Quintana, S. Schiavon, K. W. Tham, C. Miller, Balancing thermal comfort datasets: We GAN, but should
we?, in: Proceedings of the 7th ACM International Conference on Systems for Energy-Efficient Buildings,
Cities, and Transportation, BuildSys ’20, Association for Computing Machinery, New York, NY, USA,
2020, pp. 120–129.
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Occupancy is a significant area of interest within the field of building performance simulation. Through Bayesian calibration, the present study investigates the impact of the availability of different spatial resolution of occupancy data on the gap between predicted and measured energy use in buildings. The study also examines the effect of occupancy data on the quality of the constructed prediction intervals (PIs) using the Coverage Width-based Criterion (CWC) metric. CWC evaluates the PIs based on both their coverage (correctness) and width (informativeness). This investigation takes the form of an actual building case study, with nine months of hourly measured building electricity use, WiFi connection counts as a proxy for occupancy, and actual weather data. In general, the building energy model’s accuracy improves with the occupancy and plug-loads schedule derived from WiFi data. Specifically, the Coefficient of Variation Root Mean Square Error (CV[RMSE]) reduced from 37% to 24% with an exponential improvement in the PIs quality compared to the results obtained with ASHRAE 90.1 reference schedules. However, the increase in prediction accuracy shrank to 5% CV(RMSE) and a comparable CWC upon calibrating the base loads of the reference schedules. Increasing the spatial resolution from building aggregated to floor aggregated occupancy data worsened the CV(RMSE) and CWC, suggesting trade-offs between parameter uncertainty and model bias/inadequacy. These results contribute to our understanding of the interactions between model complexity, simulation objectives, and data informativeness, facilitating future discussions on the right level of abstraction when modeling occupancy.
Conference Paper
Full-text available
Thermal comfort assessment for the built environment has become more available to analysts and researchers due to the proliferation of sensors and subjective feedback methods. These data can be used for modeling comfort behavior to support design and operations towards energy efficiency and well-being. By nature, occupant subjective feedback is imbalanced as indoor conditions are designed for comfort, and responses indicating otherwise are less common. This situation creates a scenario for the machine learning workflow where class balancing as a pre-processing step might be valuable for developing predictive thermal comfort classification models with high-performance. This paper investigates the various thermal comfort dataset class balancing techniques from the literature and proposes a modified conditional Generative Adversarial Network (GAN), comfortGAN, to address this imbalance scenario. These approaches are applied to three publicly available datasets, ranging from 30 and 67 participants to a global collection of thermal comfort datasets, with 1,474; 2,067; and 66,397 data points, respectively. This work finds that a classification model trained on a balanced dataset, comprised of real and generated samples from comfortGAN, has higher performance (increase between 4% and 17% in classification accuracy) than other augmentation methods tested. However, when classes representing discomfort are merged and reduced to three, better imbalanced performance is expected, and the additional increase in performance by comfortGAN shrinks to 1-2%. These results illustrate that class balancing for thermal comfort modeling is beneficial using advanced techniques such as GANs, but its value is diminished in certain scenarios. A discussion is provided to assist potential users in determining which scenarios this process is useful and which method works best.
Full-text available
The concepts of smart building (SB) and smart facilities management (SFM) are crucial as they aim to uplift occupants' living standards through information and communication technology. However, the current network possesses several challenges to SFM, due to low bandwidth, high latency, and inability to connect a high amount of IoT (Internet of things) devices. 5G technology promises high-class network services with low latency, high bandwidth, and network slicing to achieve real-time efficiency. Moreover, 5G promises a more sustainable future as it will play a crucial role in reducing energy consumption and shaping future applications to achieve higher sustainability goals. This paper discusses the current challenges and benefits of implementing 5G in various use cases in SFM applications. Furthermore, this paper highlights the Singapore government rollout plan for 5G implementation and discusses the roadmap of SFM use case development initiatives undertaken by 5G Advanced BIM Lab (Department of Building, National University of Singapore) in alignment with the 5G implementation plan of Singapore. Under these 5G SFM projects, the lab seeks to develop state-of-the-art 5G use cases in collaboration with various industry partners and developed a framework for teaching and training to enhance students' learning motivation and help mid-career professionals to upskill and upgrade themselves to reap multiple benefits using the 5G network. This article will serve as a benchmark for researchers and industries for future progress and development of SFM systems by leveraging 5G networks for higher sustainability targets and implementing teaching and learning programs to achieve greater organizational excellence.
Full-text available
Differences in the scope and intent of the contrasting IFC and CityGML data formats entail that converting the former to the latter results in loss of information. However, for some use cases it is beneficial to keep also particular information from IFC that is not native to CityGML, and achieving that requires mechanisms such as the CityGML Application Domain Extension (ADE). We develop an ADE to support retaining relevant information from IFC. Besides being driven by the particular source of the input data (IFC), this multipurpose ADE is shaped after a discovery process that involved examining potentially applicable use cases in Singapore, doubling as an extension that is adapted to a set of use cases and the local geographic context. We implement the conceptual work by generating an enriched dataset (with an automatic conversion from IFC to CityGML), visualising it, and discuss its added value in a use case.
Conference Paper
Full-text available
As the individual difference in thermal comfort transition time is yet to be thoroughly explored, this paper studies time adaptability amongst individuals by collecting high frequency subjective comfort feedback using micro ecological momentary assessments on a smart-watch. The individuals were grouped based on their thermal preference responses and had their transition times analyzed. Each group has various transition time from arriving at and leaving from their comfort state. On average people who are more sensitive to cold temperatures, Group 1, take 8.9 minutes from being uncomfortably cold and 25.0 minutes from being uncomfortably hot to reach comfort zones. On the other hand, people who are generally comfortable, Group 2, take 22.4 minutes from uncomfortably cold and 27.1 minutes from uncomfortably hot to be thermally comfortable. The average transition time within a cluster matches the thermal comfort trend of said cluster. Ultimately, the transient time of preference groups raises the possibility to improve individualized thermal comfort models and machine learning in the future.
Full-text available
Evaluating and optimising human comfort within the built environment is challenging due to the large number of physiological, psychological and environmental variables that affect occupant comfort preference. Human perception could be helpful to capture these disparate phenomena and interpreting their impact; the challenge is collecting spatially and temporally diverse subjective feedback in a scalable way. This paper presents a methodology to collect intensive longitudinal subjective feedback of comfort-based preference using micro ecological momentary assessments on a smartwatch platform. An experiment with 30 occupants over two weeks produced 4378 field-based surveys for thermal, noise, and acoustic preference. The occupants and the spaces in which they left feedback were then clustered according to these preference tendencies. These groups were used to create different feature sets with combinations of environmental and physiological variables, for use in a multi-class classification task. These classification models were trained on a feature set that was developed from time-series attributes, environmental and near-body sensors, heart rate, and the historical preferences of both the individual and the comfort group assigned. The most accurate model had multi-class classification F1 micro scores of 64%, 80% and 86% for thermal, light, and noise preference, respectively. The discussion outlines how these models can enhance comfort preference prediction when supplementing data from installed sensors. The approach presented prompts reflection on how the building analysis community evaluates, controls, and designs indoor environments through balancing the measurement of variables with occupant preferences in an intensive longitudinal way.
Full-text available
There is an increasing activity in developing workflows and implementations to convert BIM data into CityGML. However, there are still not many platforms that are suitable to view and interact with the detailed information stored as a result of such conversions, especially if an Application Domain Extension (ADE) is involved to support additional information. We investigated the ease of use and features supported by visualisation software and tools with CityGML and ADE support, and propose an approach to develop a tool that combines useful features using a set of generic rules that can extract CityGML ADE attributes. The work, while generic, is geared towards detailed architectural datasets sourced from BIM. We implemented the approach in a web-based viewer supporting the visualisation of CityGML datasets enriched with ADE features.
Conference Paper
Full-text available
In this paper, we represent a methodology of a graph em-beddings algorithm that is used to transform labeled property graphs obtained from a Building Information Model (BIM). Industrial Foundation Classes (IFC) is a standard schema for BIM, which is utilized to convert the building data into a graph representation. We used node2Vec with biased random walks to extract semantic similarities between different building components and represent them in a multi-dimensional vector space. A case study implementation is conducted on a net-zero-energy building located at the National University of Singapore (SDE4). This approach shows promising machine learning applications in capturing the semantic relations and similarities of different building objects, more specifically, spatial and spatio-temporal data.
Full-text available
Well-being in the built environment is a topic that features frequently in building standards and certification schemes, in scholarly articles and in the general press. However, despite this surge in attention, there are still many questions on how to effectively design, measure, and nurture well-being in the built environment. Bringing together experts from academia and the building industry, this paper aims to demonstrate that the promotion of well-being requires a departure from conventional agendas. The ten questions and answers have been arranged to offer a range of perspectives on the principles and strategies that can better sustain the consideration of well-being in the design and operation of the built environment. Placing a specific focus on some of the key physical factors (e.g., light, temperature, sound, and air quality) of indoor environmental quality (IEQ) that strongly influence occupant perception of built spaces, attention is also given to the value of multi-sensory variability, to how to monitor and communicate well-being outcomes in support of organizational and operational strategies, and to future research needs and their translation into building practice and standards. Seen as a whole, a new framework emerges, accentuating the integration of diverse new competencies required to support the design and operation of built environments that respond to the multifaceted physical, physiological, and psychological needs of their occupants.
The ever-changing data science landscape is fueling innovation in the built environment context by providing new and more effective means of converting large raw data sets into value for professionals in the design, construction and operations of buildings. The literature developed due to this convergence has rapidly increased in recent years, making it difficult for traditional review approaches to cover all related papers. Therefore, this paper applies a natural language processing (NLP) method to provide an exhaustive and quantitative review.Approximately 30,000 scientific publications were retrieved from the Elsevier API to extract the relationship between data sources, data science techniques, and building energy efficiency applications across the life cycle of buildings. The text-mining and NLP analysis reveals that data sciences techniques are applied more for operation phase applications such as fault detection and diagnosis (FDD), while being under-explored in design and commissioning phases. In addition, it is pointed out that more data science techniques that are to be investigated for various applications. For example, generative adversarial networks (GANs) has potential in facilitating parametric design; transfer learning is a promising path to promoting the application of optimal building operation.