U.S. Roads: A look at our roads

This is an explorative project intented to analyze, visualize, and make sense of the data collected by the Department of Transportation's (DoT) Long Term Pavemenet Program (LTPP). LTPP is an onging program which aims to collect and research pavement data.

The Long-Term Pavement Performance Information Management System User Guide can be found here.

Additional information on Federal Highway Administratration's ongoing research efforts can be found here.

Below is a synopsis of LTPP, which this project supports, from the LTPP official brochure.

LTPP Info


A Breakdown of Our Roads

Typically our roads have several layers. The top layer is usually asphalt or cement concrete. Beneath the surface is the base and subbase, which are either concrete bound or unbound aggregates. Beneath the base/subbase is the compacted subgrade (soil).

Data Processing

The heart of this project is data analysis. In order to have meaningful and plot-able data, the team has to first sort through massive amounts of data. For example, the table which measures the representative layer thickness of various sections of highways throughout the country has over 38,000 entries, each entry with a dozen attributes.

Traffic data, likewise, has over 16,000 entries spanning over a 20 year period. To identify the road section with the most traffic, we crossed referenced yet another database containing over 2,500 entries.

Data processing was done using the Python dataframe method.

Because the scope of the data is so large, we balance analysis efforts between national average and focus studies on specific states. For obvious reasons, we are interested in the roads in Hawaii. Keep in mind the DoT manages federal highways, and there are only three such roads in the state of Hawaii: H-1, H-2 (201), and H-3. All other roads are maintained by the State and County. Since the sample size of Hawaii roads is so small, we also explored other states such as California.

Visualization

This project is for University of Hawaii at Manoa's Fall 2020 ICS 484, Data Visualization Class. Special thanks to Professor Jason Leigh for all the guidance. After obtaining data from the Department of Transportation, we used LTPP's guide book to decipher the meaning of the data. For data processing, we used Python. For Visualization, we used chart.js and plotly.js for histogram, line, map, and pie charts. Road test section visualization was done using matplotlib in Python, then translated into JS and JSON files.


Caption of data processing using Python Pandas library