Incidents on congested expressways lead to roadway capacity reduction and commuter time loss. Incident data of Shanghai inner-ring and middle-ring expressways in Pu’xi by video monitoring were employed in this study. An analysis was made of the monthly change, weekly change and daily change of incident occurrence on segments with different ramp types, and the relationship was explored between the incident occurrence and traffic flow characteristics. Negative binomial regression was adopted to establish the frequency estimation model to explore temporal heterogeneity. Results show that holiday is significant for both types of incidents while temperature only has a significant impact on disabled vehicles. The effect of weather varies among different expressways, time periods and incident types.