Digital Humanities: Digital Humanities Advancement Grants

Period of Performance

1/1/2018 - 1/31/2020

Funding Totals

$72,390.00 (approved)
$59,457.22 (awarded)

Digital Floor Plan Database: A New Method for Analyzing Architecture

FAIN: HAA-258763-18

Baylor University (Waco, TX 76798-7284)
Elise King (Project Director: June 2017 to June 2021)
King-Ip (David) Lin (Co Project Director: August 2017 to June 2021)

The continued development of a prototype of an analytical tool and database to allow humanities scholars and students to comparatively study architectural floor plans. The test case would be a collection of floor plans by American architect Frank Lloyd Wright from the Alexander Architectural Archives at the University of Texas, Austin.

Currently, those who design and study the built environment are hindered by an inability to examine large datasets of architectural drawings. Despite advancements in image recognition, no integrated system is capable of storing, reading, and analyzing floor plans. To solve this problem, this project is developing the Building Database & Analytics System (BuDAS) to partially automate the process of floor plan analysis. This project is seeking funding to expand the prototype into an integrated open source system with image recognition software for automatic floor plan detection, a database for the storage and management of data, and advanced query and graphing tools. BuDAS will allow users to compare thousands of plans to discover common design elements, examine spatial relationships over time, and mine for patterns across datasets. These findings will allow for a deeper understanding of the trends and patterns of space usage and the relationship between buildings and human experience.

Associated Products

The Building Database and Analytics System (BuDAS): Computer science and interior design take on ‘big data' (Conference Paper/Presentation)
Title: The Building Database and Analytics System (BuDAS): Computer science and interior design take on ‘big data'
Author: Elise King
Author: Qiannan Wu
Author: David Lin
Abstract: Mass digitization has allowed greater access to archival materials than ever before. While access has certainly improved, processing and analyzing the rich information contained in these drawings remains challenging. In the age of ‘big data,’ having access to information is no longer enough; we need tools to sift through massive amounts of raw information to identify meaning and pattern (Manovich, 2012). Frustrated by the lack of tools to analyze archival floor plans, the authors considered the following question: How can we automate the collection of floor plan information and explore patterns and relationships between plans, architects/designers, and time periods? In this paper we explore the challenges of analyzing archival floor plans and provide an overview of the Building Database & Analytics System (BuDAS), which we developed to address these problems. (abbreviated abstract)
Date: 10/03/19
Primary URL:
Conference Name: Interior Design Educators Council Southwest Regional Conference


Best Paper
Date: 10/3/2019
Organization: Interior Design Educators Council Southwest Regional Conference

Building Database Analytics System (BuDAS): Examining challenges to floor plan detection (Conference Paper/Presentation)
Title: Building Database Analytics System (BuDAS): Examining challenges to floor plan detection
Author: Elise King
Author: David Lin
Abstract: The success of plan recognition is determined by two main factors: 1). quality of the plan image, and 2). interpretation of the plan information. Plan image quality includes factors such as clarity, contrast, and digital noise. A high-quality plan image has high contrast, minimal noise, and clear lines and text. Plan interpretation is essentially how well the system is able to process and make sense of the layers of plan information. A basic floor plan that includes walls, windows, doors, and room labels is easier for the system to interpret than a plan with additional layers of information, such as furniture, dimension stringers, material symbols, grids, landscaping, and ceilings or rooflines. These additional layers of information can result in the plan detection misidentifying other lines as walls, for example. By testing each of these variables, we are able to determine how to improve plan for users. This poster will explore the results of these tests and suggest areas for improvement. (abbreviated abstract)
Date: 10/26/19
Conference Name: Digital Frontiers Conference

Exploring Frank Lloyd Wright through the Lens of ‘Big Data’ (Conference Paper/Presentation)
Title: Exploring Frank Lloyd Wright through the Lens of ‘Big Data’
Author: Elise King
Author: Qiannan Wu
Author: David Lin
Abstract: Frank Lloyd Wright remains one of America’s best known and most prolific architects, credited with designing more than 1,000 structures. Such a sizable oeuvre holds hundreds of thousands of data points. Much of this data, however, remains untapped. Wright’s floor plans, for example, are full of rich layers of information (e.g., tracking room names over time can indicate changing societal preferences/shifts). But to utilize the data in hundreds or thousands of these plans, requires either reproducing them in CAD or BIM or measuring and recording the information by hand. Both are time consuming and prone to error. In response, we developed a tool that allows users to analyze large corpora of architectural floor plan images. The Building Database & Analytics System (BuDAS) unites complementary manual data entry and image recognition to automate the process of floor plan detection and analysis. In this study, we explore Wright’s changing conception of space through an analysis of his residential floor plans using BuDAS. Space is a topic Wright discussed frequently throughout his career, including in his autobiography and principles of organic architecture, among other venues. But do Wright’s pontifications on space stand in accordance with his built work? Using floor plan images of Wright’s residences uploaded to BuDAS, we collected information on room relationships, doors/openings, room sizes, and room names/categories. After uploading plan images, BuDAS automatically detects and collects information related to room sizes, types, and relationships. This, along with contextual information entered manually (e.g., location, date of construction, cost, name of client), is stored in the database for analysis. Using the data collected from Wright’s residential floor plans, we explore interior and exterior space usage (minimum, average, and maximum square footage) and room relationships (connectivity, adjacency, and openness) over the course of his career.
Date: 03/06/2020
Primary URL:

Building Database and Analytics System (BuDAS) (Computer Program)
Title: Building Database and Analytics System (BuDAS)
Author: Elise King
Author: David Lin
Abstract: We are committed to the creation of open and accessible tools. The extractor, annotator, and database are available through the project website ( or Github ( and They are implemented in Python 3.7 using publicly available modules. Our team chose Python because it is commonly used in digital humanities and offers flexibility and cross-compatibility. We also included an executable option, but it is only available for Windows users at this time. Our team created documentation and instructions for all the above components, which are available through the Github links.
Year: 2019
Primary URL:
Primary URL Description: BuDAS extractor and annotator
Secondary URL:
Secondary URL Description: BuDAS database and analyzer
Access Model: Open access
Programming Language/Platform: Python
Source Available?: Yes

BuDAS Project Website (Web Resource)
Title: BuDAS Project Website
Author: Elise King
Author: David Lin
Abstract: BuDAS project website to provide basic information about the project and link to Github and other resources.
Year: 2019
Primary URL: