We analyze the evolution of Fe XII coronal plasma upflows from the edges of ten active regions (ARs) as they cross the solar disk using the Hinode Extreme Ultraviolet Imaging Spectrometer (EIS). Confirming the results of Demoulin et al. (2013, Sol. Phys. 283, 341), we find that for each AR there is an observed long term evolution of the upflows which is largely due to the solar rotation progressively changing the viewpoint of dominantly stationary upflows. From this projection effect, we estimate the unprojected upflow velocity and its inclination to the local vertical. AR upflows typically fan away from the AR core by 40 deg. to near vertical for the following polarity. The span of inclination angles is more spread for the leading polarity with flows angled from -29 deg. (inclined towards the AR center) to 28 deg. (directed away from the AR). In addition to the limb-to-limb apparent evolution, we identify an intrinsic evolution of the upflows due to coronal activity which is AR dependent. Further, line widths are correlated with Doppler velocities only for the few ARs having the largest velocities. We conclude that for the line widths to be affected by the solar rotation, the spatial gradient of the upflow velocities must be large enough such that the line broadening exceeds the thermal line width of Fe XII. Finally, we find that upflows occurring in pairs or multiple pairs is a common feature of ARs observed by Hinode/EIS, with up to four pairs present in AR 11575. This is important for constraining the upflow driving mechanism as it implies that the mechanism is not a local one occurring over a single polarity. AR upflows originating from reconnection along quasi-separatrix layers (QSLs) between over-pressure AR loops and neighboring under-pressure loops is consistent with upflows occurring in pairs, unlike other proposed mechanisms acting locally in one polarity.