The Iowa Department of Transportation (Iowa DOT) has made it a priority to address mobility and safety by identifying traffic critical work zones across the state and deploying various mitigation strategies. This paper presents a case-study on the Iowa DOT’s Traffic Critical Project program aiming to get a uniform, transparent and systematic performance visualization of on-going work zone projects based on the streaming data collected from radar sensors. These performance visualizations both help the DOT monitor traffic behavior in work zones on a daily basis as well as assist in the decision making to improve work zone performance. To accommodate the large amount of traffic surveillance data and perform analysis on the big data sets in a timely manner, this visualization system integrates several big-data techniques. Hadoop distributed file system (HDFS) is used to provide a large-capacity and fault-tolerant data storage as well as a distributed computing environment. The data processing speed of major tasks such as filtering, grouping and aggregating is boosted up by distributed computing under MapReduce framework. The pre-processed data are finally converted to a column-oriented database which optimizes data reading speed to support interactive data queries. Tableau workbooks are used to design the graphical visualization of the traffic characteristics and work zone performance measures. Then a web-based interface is created for safe and convenient access. This big-data-driven traffic surveillance system provides the Iowa DOT a tool to understand the sensor working condition and the impact the Traffic Critical Project (TCP) work zones have on traffic and the decision making to improve mobility and safety.