Open Plans

Open Plans is an online, open and collaborative platform addressing the challenge of accessing and utilizing the wealth of architectural knowledge embedded in building floor plans. The motivation for the project was to provide architects, researchers and students with an intuitive, efficient and comprehensive tool to unlock the vast potential of architectural plans for research and design purposes. Plans play a crucial role in architecture, as they are essential for architects to understand, communicate, and realize buildings. Plans encode the structure and organization of a building, but this knowledge is implicitly embedded and difficult to harvest with conventional approaches. Despite being abundant in books, archives, and raster images, the architecturally significant features stored in plans remain inaccessible to data-driven approaches and research. The motivation behind Open Plans is to leverage the latest data processing methods specific to the discipline and enable access to the vast architectural knowledge embedded in floor plans. A key aspect of the project is the creation of an expandable database of semantically labeled floor plans. A web-based graphical interface allows users to add not only project and plan information but also details about the scale, orientation, form, functions, and qualities of a building. By building such a large and flexible database, the project provides a basis for AI and data-driven approaches for architectural research. The possibilities include gaining new insights by discovering lesser-known patterns of buildings or exploring augmented design tools based on the references of already existing houses. Open Plans fosters a collaborative and open environment by providing each user with a dedicated user profile. Users can conveniently access their contributions under ‹My Projects›, showing their input to the platform. Diverse research groups have the opportunity to contribute their specific expertise to the expanding database. By encouraging collaboration and knowledge sharing, the platform aims to create a dynamic and comprehensive resource that benefits the architectural community. Open Plans is accessible online via plans.arch.ethz.ch. The platform offers various functionalities for retrieving information, including the ‹Search Plans by Drawing› feature, enabling users to explore the database by sketching a building outline, and the ‹Projects Map› functionality, facilitating navigation to specific geo-locations. These architecture-specific queries provide an intuitive means of accessing information. Overall, the project bridges the gap between the abundant resources of building floor plans and the latest data-driven approaches to enhance architectural research and design.

Open Plans is integrated into the Computational Design I-II course, where 350 Bachelor students engage with the platform and continue to have access to it as a valuable repository of built references. The students are introduced to the Open Plans platform through a lecture, video tutorials, and flipped classroom sessions, where they learn to navigate the platform and search for architectural plans. A Grasshopper component allows students to connect with familiar CAD environments like Rhino3D + Grasshopper and to interact with the Open Plans database directly. The teaching implementation of Open Plans includes an assignment that requires students to contribute plans to the growing database. They are tasked with sourcing exemplary residential buildings from the library, translating them into inputs for Open Plans via the online interface, and providing detailed descriptions of each plan. Each submission enriches the database, enabling others to explore and learn from a diverse range of design approaches, spatial arrangements, and functional solutions. In a second assignment, students are tasked with querying the database using Grasshopper, extracting relevant data, and utilizing it to create compelling data visualizations. The final outcome of this task is a series of posters that demonstrate the potential of harnessing simple yet effective data-driven techniques to communicate concepts visually. Through this teaching implementation, Open Plans not only serves as a valuable resource for students to explore architectural knowledge but also actively engages them in contributing to and benefiting from the collective repository of plans. This hands-on experience broadened their understanding of architectural concepts and enhanced their skills in data-driven approaches and design visualization.

Open Plans has been successfully implemented as an accessible online platform, with an intuitive graphical interface that allowed 350 Bachelor students to upload 3500 new plans and over 8000 semantic labels. The main achievement of the project lies in the development of an openly available database that can serve as an invaluable resource for architectural research and education. Documentation of the project on GitBook offers support materials for larger student populations and researchers interested in integrating Open Plans into their teaching or research practices. To evaluate the influence of the project on student learning, a feedback session was conducted to collect insights on the usability of the Open Plans platform. Students found the platform to be intuitive and expressed enthusiasm about their contributions becoming part of the collective repository of plans. This suggests Open Plans offered students a hands-on learning experience on the uploading, querying, and organization of data. Throughout the assignment phase, it became evident that slight differences in data gathering had a notable influence on the overall quality of the database. To address this, additional development is required to enable the rapid expansion of the database, particularly when it becomes accessible to larger student bodies or other organizations. Many functional extensions are possible based on the current development: for example, extending the functionality of Open Plans to “Playlists” to create custom collections would be a meaningful next step. Establishing partnerships with other institutions is an exciting outlook for the future of Open Plans, as it allows for the platform’s functionalities to be enhanced by incorporating contributions from other researchers. Showcasing the use of the Open Plans database for machine learning approaches in architectural research has the potential to unlock new insights and patterns, which could be an exciting development for architectural research.