■  Data Analytics of Accidents and Polices using Python, R

  Under the Chemical Substances Control Act of South Korea, it is mandated to submit transportation plans containing information on the transport of hazardous chemicals, with over 600,000 submissions recorded annually. An big data analysis was conducted on 2,506,985 transportation plans to discern trends and assess their correlation with chemical transport accidents.

  A total of 798 chemical accident records reported in South Korea were obtained from the Korean Ministry of Environment and subjected to detailed analysis. The data set consists of a total of 18 variables including accident region, cause of accident, accident type, and number of casualties. A data set in the form of a 798 × 18 matrix was constructed, and correlation analysis was performed using methods appropriate to the type of each variable.

  Assessing the impact of internal and external environmental changes on systems related to hazardous chemical handling facilities is a crucial aspect of future system improvements. Among the notices concerning handling facilities, civil complaints related to small quantity standards and small quantity handling facilities were the highest. Inquiries regarding damage reduction facility standards, such as bund walls and detectors, ranked prominently, suggesting a need for policy service improvements in this area.

  A predictive model was developed to identify workplaces at high risk of hazardous chemical accidents. The analysis focused on examining the relationships between aging industrial complexes and facility defects, noncompliance with safety standards and incomplete safety training, as well as the spatial distribution of accident sites in relation to complaint records documented in the Chemical Accident Prevention and Management Plan. Key variables associated with the primary causes of accidents were identified through this analysis, forming the basis for the development of the accident prediction model.