Consistent efforts with dense sensor deployment and data gathering processes for bridge big data have accumulated profound information regarding bridge performance, associated environments, and traffic flows. However, direct applications of bridge big data to long-term decision-making processes are hampered by big data-related challenges, including the immense size and volume of datasets, too many variables, heterogeneous data types, and, most importantly, missing data. The objective of this project was to develop a foundational computational framework that can facilitate data collection, data squashing, data merging, data curing, and, ultimately, data prediction. By using the framework, practitioners and researchers can learn from past data, predict various information regarding long-term bridge performance, and conduct data-driven efficient planning for bridge management and improvement. This research project developed and validated several computational tools for the aforementioned objectives. The programs include (1) a data-squashing tool that can shrink years-long bridge strain sensor data to manageable datasets, (2) a data-merging tool that can synchronize bridge strain sensor data and traffic flow sensor data, (3) a data-curing framework that can fill in arbitrarily missing data with statistically reliable values, and (4) a data-prediction tool that can accurately predict bridge and traffic data. In tandem, this project performed a foundational investigation into dense surface sensors, which will serve as a new data source in the near future. The resultant hybrid datasets, detailed manuals, and examples of all programs have been developed and are shared via web folders. The conclusion from this research was that the developed framework will serve practitioners and researchers as a powerful tool for making big data-driven predictions regarding the long-term behavior of bridges and relevant traffic information.