Understanding relationships between stream chemistry and watershed factors: land use/land cover, climate, and lithology are crucial to improving our knowledge of critical zone processes that influence water quality. We compiled major ion data from more than 100 monitoring stations collected over 60 years (1958-2018) across the Colorado River Watershed in Texas (103,000 km²). We paired this river chemistry data with complementary lithology, land use, climate and stream discharge information. A combination of graphical geochemistry and machine learning techniques were used to produce new insights on controls of stream water chemical behavior. Studies on stream flow and chemistry in the American west and globally have shown strong relationships between major ion chemical composition and lithology, which hold true for the Colorado River basin in this study. Reactive minerals, including carbonates and evaporites, dominate major ion chemistry across the upper watershed. Upstream and central reaches of the Colorado River showed shifts from Na-Cl-SO₄ dominated water from multiple sources including dissolution of gypsum and halite in shallow groundwater, agricultural activities, and oil and gas development, to Ca-HCO₃ water types controlled by carbonate dissolution. In the lower portion of the watershed multiple analyses demonstrate that stream chemistry is more influenced by greater precipitation and the presence of relatively fewer reactive silicate minerals than middle and upstream reaches. This study demonstrates the power of applying machine learning approaches to publicly available long term water chemistry datasets to improve the understanding of water and nutrient cycling, salinity sources, and water use.