Abstract: Cartograms is a new approach for statistical data visualization.However,drastic diversity oflocal data leads to failures when we apply existing cartogram-making algorithms on grid statistic datawithin urban scale.To address this problem,we design an index for measuring the scale effect of gridstatistic data.The scale effect index well describes the restriction effect and representation effect of1104武汉大学学报·信息科学版2015年8月 multi-scale grid data.Moreover,based on the scale effect index,we modify the diffusion-based meth-od developed by Gastner and Newman in 2004.The modified algorithm is not sensitive to local diversi-ties of grid statistic data,and area cartograms achieved by this algorithm well keep the density-equali-zing condition.We apply this algorithm in making gridded population cartograms from 100mgrid Bei-jing census data,and results show the reliability of the algorithm,which can be used for producing ar-ea cartograms of statistic data.