1 of 1

The Internet is composed of 50 billion websites and grows larger every day. As the number of links and specialty subject areas increases, it becomes more difficult to find pertinent information. Special purpose data crawlers continually search the Internet for specific information on, for example, real estate, air travel, auto sales, and others. Such single-purpose crawlers can search for hundreds of keywords and use machine learning to determine whether what is found is relevant. The use of such special purpose data crawlers and associated knowledge databases also allows the collection and analysis of agricultural and forestry data. In this presentation, we discuss the development, architecture, functioning, and applications of such specialty knowledge database and crawler system to find information related to two growing forest products sectors: Urban & Reclaimed Wood Utilization, and Mass Timber.

​

Urban and Reclaimed Wood: There are approximately 74 billion existing urban trees, and when these need to be removed or pruned, they are considered “wood waste.” In 2010, it was estimated that of the 34.2 million tons of wood-based municipal solid waste (MSW) generated, 18.4 million tons of woody yard trimmings like urban trees and limbs were disposed of, with 4.0 million tons available for recovery (Figure 1)^[1]. For the purpose of this project, reclaimed wood consist of all previously utilized wood products brought back into circulation, largely originating from structures like old barns and buildings (Figure 2). It is estimated that approximately 36.4 million tons of C&D wood waste are generated annually, with approximately 17.3 million tons available for recovery^[1][2]. To add additional value, industries have emerged to provide an outlet for these sources, foster local economies, and divert supply to landfills (Figure 1). Beyond environmental and social benefits, value-added urban and reclaimed wood products provide consumers with unique aesthetics, historical significance, and sentimentality derived from weathered and inimitable wood supplies.

​

​

​

​

​

​

​

​

​

Mass Timber is a family of engineered wood products, which consist mostly of premanufactured and multilayered solid wood structural elements. Mass timber includes Cross Laminated Timber, made of orthogonally bonded layers of solid lumber, laminated by gluing of longitudinal and transverse layers, to form a solid rectangular-shaped element intended for roof, floor, or wall applications³. Mass timber buildings as a structural system allows for short erection times and lower costs compared to steel and concrete, with relatively little waste, and less disturbance to a construction site’s surroundings. CLT has also environmental benefits compared to other structural systems.

​

​

​

​

​

​

​

Department of Bioproducts and

Biosystems Engineering

​

​

Development of Knowledge Databases to Support US Forest Products:

urbanwooddatabase.umn.edu & masstimberdatabase.umn.edu

Introduction

Omar Espinoza, Associate Professor, U of M

Ed Thomas, Research Computer Scientist, US Forest Service

Methods

Outcomes

To achieve the objectives of this research, the project was executed in three main stages: development of system objectives, requirements, and capabilities; system development; system testing; and dissemination. The system’s architecture is illustrated in Figure 3.

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

The objective of this project was to develop a specialized search engine and knowledge databases that operates and collects relevant links and data related to two emerging forest products sector, with minimal human oversight. The most important requirement was the ability to keyword-search websites and web documents for potentially hundreds of keywords, categorize information found, and store those links in a knowledge database for easy and fast retrieval later. Hence, the crawler was tasked to search not only web pages but also all kinds of documents referred to on these websites. Such documents are posted in countless formats, including, but not restricted to, Portable Document Format (PDF), text documents, presentations, and others, requiring the crawler to be able to process a wide array of file formats. Each source needs to be searched for the keywords, all relevant documents need to be stored in the knowledge database and, most importantly, the system needs to find and show the relevant knowledge upon the users entering a search term in an easy-to-use interface.

​

To assure relevance and timeliness, the knowledge database system operates multiple crawlers at once to build the database, maintain and verify links and knowledge. However, finding and maintaining knowledge is just one critically important activity. Just as important is the ability to assess the relevance of the knowledge found and to categorize it according to the system developed, in a number of different sub-topics established by the research team (Table 1). Finally, documents that are added to the system’s knowledge database can be manually checked by an administrator before they are added to the database.

​

​

​

​

​

​

​

​

​

​

​

​

​

​

For the two databases described here, Table 2 lists the most common topics found for both databases, and Table 3 includes the resources types in the live records.

​

​

Table 1: Subject categories for the organization of the two databases

References

1. Howe, J., Bowyer, J., Pepke, E., Frank, M., & Fernholz, K. (2014). Municipal Solid Waste (MSW) and Construction and Demolition (C&D) Wood Waste Generation Recovery in the United States.
2. Howe, J., Bratkovich, S., Bowyer, J., Frank, M., & Fernholz, K. (2013). The Current State of Wood Reuse and Recycling in North America and Reccommendations for Improvements.
3. Karacabeyli, E. and B. Douglas, eds. Cross-Laminated Timber Handbook. ed. E. Karacabeyli and B. Douglas. 2013, FPInnovations and Binational Softwood Lumber Council: Pointe-Claire, QC. 572.
1. Photo Source: Ana Pitti
2. Photo Source: Omar Espinoza
3. Photo Source: Maria Fernanda Laguarda Mallo

Table 3: Counts of resources discovered by type in the urban wood database (left) and the mas timber database (right).

Figure 4: Screenshot of the user interface for the urban and reclaimed wood database.

Figure 3: Database's system architecture.

The User Interface (Figure 4) was designed with simplicity in mind. Efforts were made to make the search and filtering tools as intuitive as possible. Search, filtering, and sorting actions include text search, filtering by category (topic) and/or type of link and sorting by year. When no specific sorting criteria is selected, links displayed are sorted by the date they were added to the database.

Conclusions

Knowledge database systems foster awareness of urban and reclaimed wood, and mass timber products and applications, by disseminating knowledge and facilitating collaboration among stakeholders, while, at the same time, reducing the risk of duplication of efforts. By improving the availability of information, we believe that manufactures, suppliers, researchers, design professionals, government agencies, and other stakeholders can directly benefit from the tools presented here.

​

The work on which this poster is based was funded in whole or in part through a grant awarded by the USDA Forest Service.

Table 2: Top 10 most common sub-topics found in the urban wood database (left) and the mas timber database (right).

Objectives

The main objective of this project was to develop a knowledge management system (KMS) to support the development of the forest products sector.