Pedestrian exposure is a necessary component for a meaningful evaluation of pedestrian safety. The Space Syntax approach has a track record of accurate prediction of pedestrian activity by estimating the physical street connectivity in urban environments. However, for some environments, the performance of Space Syntax is limited and cannot be used as a reliable estimate of exposure. This paper makes use of the interdependency between: (i) street connectivity-estimated here using integration; and (ii) land-use characteristics; to propose a mechanism to adjust integration by land-use features at the block level. Different levels of integration for each street-block, which hold the same mean values along the same street, are weighted based on dominant land-use features. The weighted integration value for a street-block dominated by commercial property is higher than the mean integration value for that street. Conversely, the weighted integration value for a residential street-block is lower than the mean integration value for that street. The proposed approach captures the heterogeneity of street-blocks, which is not always captured by Space Syntax. Applying this method to the northern periphery of the University of California, Berkeley, has produced promising preliminary results. It was shown that the weighted integration values (at the street-block level) are better correlated with pedestrian volumes than mean integration values (street scale). Further research efforts are required to develop this simplified approach into a pedestrian exposure prediction model.