Abstract: This paper presents a new squarified treemap layout algorithm by leveraging mean dichotomy driven blockwise sorting. The input dataset is firstly divided into several subsets by means of the mean dichotomy technique concerning the number of allowed disordered rectangles. Then the drawing area is segmented into pieces according to the total weight of each subset. Each piece is sequentially assigned to its corresponding subset by considering its weight, and is located using the squarified self-adaptive algorithm with the long-side place strategy. Compared with previous approaches, our approach effectively reduces the time of segmentation and sorting, and addresses several issues remained in data segmentation. In addition, ours maintains the order of the input data elements and improves the layout stability.