Examining the maintenance patterns across regional levels using statistical analysis on bridge inspection records
Presenter Type
UNO Graduate Student (Masters)
Other
Data Science
Advisor Information
Professor
Location
MBSC302 - G (Masters)
Presentation Type
Oral Presentation
Start Date
24-3-2023 10:30 AM
End Date
4-3-2023 11:45 AM
Abstract
In the U.S., there are over 600,000 bridges that serve as the backbone of transportation infrastructure in the U.S. The 2021 Infrastructure report card grades C for the condition of the bridges. The average age of the bridge is around 50 years, the same as the design life of the bridge. Previous research studies have adopted statistical analysis and machine learning models to identify influential variables contributing to the deterioration and maintenance of bridges. Understanding how bridges deteriorate and maintain will guide us in making high-stakes decisions.
Features like age, average daily traffic, maintenance, and material are often reported as the most influential variables. However, many of these results were studied at national and state levels. Whereas every state in the U.S. has its own Department of Transportation (DoT) that maintains and follows its guidelines. Further, recent survey results suggest that county owners might have intervention guidelines in managing the bridges within these states. As a result, the influential rules and patterns from the modeling at the state level might need to be more generalized. There needs to be more understanding of the granularity at which the machine learning model should be trained for predicting bridge deterioration and maintenance. Moreover, there needs to be more understanding of the generalized rules and patterns observed from modeling at that the state-level rules are applicable at the county level as well. Also, previous literature reported the individual contribution of each variable. This research study aims at identifying influential factors and how the influential factors interact in explaining bridge maintenance to address the gaps in the literature. The analysis is based on the National Bridge Inventory (NBI) dataset from 1992 to 2022 for Nebraska. This research adopts a comprehensive exploratory method to analyze essential factors independent of machine learning models to validate the previous understanding of bridge maintenance. Further, the analysis will consider the granularity of data to analyze the level of state, county, and specific region. This study extends its analysis to test the most reported factors in the literature of bridge deterioration and maintenance for all sub-components of the bridges. The study adopts ANOVA to analyze the factors within the NBI dataset from 1992 to 2022 that influence or exhibit a pattern concerning bridge maintenance. Further, this analysis is carried out on all sub-components of the bridges.
Scheduling
10:45 a.m.-Noon
Examining the maintenance patterns across regional levels using statistical analysis on bridge inspection records
MBSC302 - G (Masters)
In the U.S., there are over 600,000 bridges that serve as the backbone of transportation infrastructure in the U.S. The 2021 Infrastructure report card grades C for the condition of the bridges. The average age of the bridge is around 50 years, the same as the design life of the bridge. Previous research studies have adopted statistical analysis and machine learning models to identify influential variables contributing to the deterioration and maintenance of bridges. Understanding how bridges deteriorate and maintain will guide us in making high-stakes decisions.
Features like age, average daily traffic, maintenance, and material are often reported as the most influential variables. However, many of these results were studied at national and state levels. Whereas every state in the U.S. has its own Department of Transportation (DoT) that maintains and follows its guidelines. Further, recent survey results suggest that county owners might have intervention guidelines in managing the bridges within these states. As a result, the influential rules and patterns from the modeling at the state level might need to be more generalized. There needs to be more understanding of the granularity at which the machine learning model should be trained for predicting bridge deterioration and maintenance. Moreover, there needs to be more understanding of the generalized rules and patterns observed from modeling at that the state-level rules are applicable at the county level as well. Also, previous literature reported the individual contribution of each variable. This research study aims at identifying influential factors and how the influential factors interact in explaining bridge maintenance to address the gaps in the literature. The analysis is based on the National Bridge Inventory (NBI) dataset from 1992 to 2022 for Nebraska. This research adopts a comprehensive exploratory method to analyze essential factors independent of machine learning models to validate the previous understanding of bridge maintenance. Further, the analysis will consider the granularity of data to analyze the level of state, county, and specific region. This study extends its analysis to test the most reported factors in the literature of bridge deterioration and maintenance for all sub-components of the bridges. The study adopts ANOVA to analyze the factors within the NBI dataset from 1992 to 2022 that influence or exhibit a pattern concerning bridge maintenance. Further, this analysis is carried out on all sub-components of the bridges.