There is a df_new Pandas dataframe: id route operator group_name district injury_result incident_type victim_category victim_age 2 3.0 1 London General Go-Ahead Southwark NaN Onboard Injuries 7 10 3 4.0 4 Metroline Metroline Islington NaN Onboard Injuries 7 2 4 5.0 5 East London Stagecoach Havering NaN Onboard Injuries 7 8 5 6.0 5 East London Stagecoach None London Borough NaN Onboard Injuries 7 8 Count which routes had the most incidents_type

from collections import Counter

def most_incident(df, column):
    # Count incident_type by route
    incident_id = Counter(df[column])
    # Return key of the most incident_type
    return incident_id.most_common(1)[0][0]

most_incident(df_new, 'route')