We present Burger Charts, a quantitative display of set intersections that stack the contributing sets on top of each other. Sets are represented as horizontal layers. The size of each set intersection is encoded by the width of its layer section, forming a vertically stacked, burger-like visual representation. A visual skewer maintains the unity of the burger by bridging gaps of those set layers that do not contribute to the intersection. The color coding of the sets emphasizes which set contributes to which intersection. We use Burger Charts to visualize and analyze keyword co-occurrences of an argument search engine. They support a quantitative discourse analysis by providing insights into the distribution of sets of keyword occurrences. Users can interactively explore keywords of online discourses on controversial topics, identify prevalent keyword co-occurrences, and even uncover overlooked perspectives.