Nigerian Multi-modal Road Traffic Crash Data

This is a Nigerian multi-modal road traffic crash (RTC) dataset curated from the Nigerian RTC corpus collected by Idakwo et al. (2025).   The dataset has 59 features across 3 modes  as follows: Mode 1: unstructured textual data in column 'content_clean' Mode 2: structured numerical data in columns 'day_of_week', 'rtc_time', 'casualty', 'injured', 'no_vehicles', 'lost_control', 'OVL', 'headon_collision', 'rear_side_collision', 'avoid_collision', 'detached_container', 'SPV', 'BRD', 'WOT_DOT', 'OBS', 'BFL', 'TBT', 'slope', 'off_road', 'water_body', 'RTV', 'DGD', 'spillage', 'rain', 'poor_visibility', 'animals', 'non_inflammable_goods', 'inflammable_goods', 'fire', 'somersaulted', 'unregistered_vehicle', 'checkpoint', 'pedestrian', 'articulated', 'motorcycle', 'tricycle', 'car', 'bus', 'buildings', 'MDV', 'FTQ_SOS', 'UPWD', 'DAD', 'SLV', 'female', 'children', 'male', 'PWR', 'OTH', 'elderly', 'junction', 'fly over', 'bend', 'roundabout', 'turning', 'construction_zone' Mode 3: multi-modal (fused Modes 1 and 2) data in columns           'new_fused_1_2' The last column 'severity' is the target/class column. Some of the abbreviated RTC causative factors and their meanings are: ‘SPV’ - speed violation; ‘UPWD’ - use of phone while driving; ‘TBT’ - tyre burst; ‘MDV’ - mechanically deficient vehicle; ‘BFL’ - brake failure; ‘OVL‘ - overloading; ‘WOT_DOT‘ - wrongful or dangerous overtaking; ‘DGD’ - dangerous driving; ‘BRD’ - bad road; ‘RTV’ - route violation; ‘OBS’ - road obstruction violation; ‘DAD’ - driving under alcohol/drug influence; ‘PWR’ - poor weather; ’FTQ_SOS’ -  fatigue or sleeping on steering; ‘SLV’ - sign light violation; and ‘OTH’ - others (avoid collision and detached container).