Decisions towards improvement of road safety are strongly based on accident data, and data of the vehicles and victims involved in accidents. Non-availability of these data or false data can lead to wrong estimates and planning, making road safety exercise ineffective. The present study reports the parameters collected in road traffic accidents data from five mid-sized Indian cities namely Agra, Amritsar, Bhopal, Ludhiana and Visakhapatnam. Relative to road traffic crash, the data is captured focusing details of all relevant parameters of all dimensions in accident covering participating vehicle, impacted victims, pedestrian and accident related details. The data collection technique for traffic crashes across all the cities are similar but on an average form 20% up to 60% of data are not being recorded. The observation at city level data is extracted and cleaned from raw information and found that it shows a different kind of mixed pattern for recorded data and data not recorded (or missing data). Further, it presents the various classifications of missing data formats. Analysis of the data patterns gives a quantum of missing data and enables right categorization of the missing data. It further provides suggestions on various possibilities of filling the gaps for the missing data which can be used in further decision making. The data thus categorized will be used for designing a better traffic safety system.