Most modern DBMS use query optimization techniques to accelerate query execution by generating efficient plans for queries on large datasets. However, many data-intensive applications like stream processing, time series analysis, and data visualization (that do not use a DBMS as the underlying data storage framework) lack mechanisms for query optimization. To address this gap, Zhang et al. extend an OSS compiler (GraaIVM compiler and Truffle) with database-style query optimization. Upon evaluation of this system using the TPC-H benchmark, weather visualization, and microbenchmark queries, they find that their systems can choose the best plan for the current data, including by dynamically responding to changes in the data distribution.