As in any industry exposed to risk, human and organizational factors are the main stakes of maritime safety. Understanding the causes and risks of maritime accidents is integral to the sustainability of shipping. The investigation of marine accidents is a crucial tool for their identification in areas related to operations and ships, including social and situational systems, their design, and technical systems. The authors conducted a cause–effect analysis of marine incidents. For this purpose, case-by-case analysis and an Ishikawa diagram were used, which is a tool that helps identify actual or potential causes of accidents. The study showed that by far the most significant cross-section of causes of accidents were elements of social and situational systems that affect the safety of the ship, crew, and environment. The least significant contribution came from the machinery area. Through the detailed descriptions, a picture emerges not so much of a lack of knowledge of the regulations as of a failure to comply with existing procedures or best practices. In the authors’ opinion, more emphasis is needed on preventive measures, including safety culture, training, competence assessment, and increased awareness of the need for sustainability.

The article focuses on the transport preferences of the inhabitants of a medium-sized city in the context of changing urban mobility and sustainable development. It fills the gap in the literature researching the changing communication behavior of the inhabitants of a given city. The main purpose of this article is to identify and analyze the preferred means of movement for inhabitants of a medium-sized city in Central and Eastern Europe. The research area was the city of Szczecin, Poland. The research was carried out using the following research methods: literature review, analysis of legal acts and documents, case study. The research material was collected using an anonymous questionnaire based on a proprietary survey made available to the inhabitants of the largest city of north-western Poland, Szczecin, in the GOOGLE browser in the period 17 May 2019–1 April 2021. The multiple-response contingency tables, chi-square test of independence and non-parametric U Mann–Whitney and Kruskal–Wallis tests were used for study development. The conducted scientific research showed that, when determining the relationship between the method and the destination, the respondents most often chose a car or on-foot commuting in all of the surveyed destinations. The research results increase the level of knowledge regarding residents’ needs in terms of urban mobility and the development of Szczecin’s transport system. They can be used in the rational actions of local authorities aiming to improve residents’ the quality of life in the context of the sustainable development of city transport.

The need for urban societies to move continues to grow with the intensity of their various activities. One of the challenges in micromobility in cities based on bike, scooter, or public scooter systems is determining the potential yet realistic demand for such services. This article aims to present the level of demand for eco-friendly non-motorized transport systems and identify the motivations driving users of public transport to choose these eco-friendly Personal Mobility Devices (PMD). The authors propose a methodology based on the integration of four components: a case study, an analysis of participatory budget projects, a metadata analysis of bikesharing (BSS) in Szczecin, and a meta plan. They utilized a case study method and conducted a survey based on a custom questionnaire. An analysis of participatory budget projects and data from the operation of bikesharing in Szczecin was performed using statistical methods. Applying the heuristic tool meta plan, they outlined the directions for the development and support of bikesharing as an example of micromobility in Szczecin. The research identified connections between the socio-demographic structure of respondents and attributes of bike communication. This study highlighted periods of intensive system use and locations with the highest exploitation. It showed that the demand for BSS depends on weather conditions, traveler convenience, the density and coherence of bike infrastructure with important locations for residents, transfer connections, and safety. The presented research results can assist planners and authorities in Szczecin and other cities in expanding bikesharing.

Purpose: The aim of the article is to examine the possibility of using artificial neural networks to predict the concentration of PM10 from road transport on the example of the urban agglomeration in Poland, Szczecin. Design/methodology/approach: Forecasting was carried out using measurement data for the concentration of particulate matter PM10 and meteorological parameters in the form of average daily air temperature, average daily atmospheric pressure, average daily relative humidity, average daily wind speed and daily precipitation (rainfall and snowfall). The data covered the period from January 1, 2020 to December 31, 2023. For the predictive analysis, NARX neural networks with the Levenberg-Marquardt gradient algorithm with four different architectures were used (5-5-1, 5-10-1, 5-20-1, 5-50-1), which differed in the number of neurons in the hidden layer. The Pearson’s linear correlation coefficient (R) and mean squared error (MSE), were used to assess the models. Findings: It was found that architecture 5-20-1 with 20 neurons in the hidden layer obtained the highest value of correlation coefficient for overall regression, equal to 0.7379. At the same time, the 5-20-1 architecture achieved also the best performance in the learning process, obtaining the lowest MSE value of validation equal to 0.00596. Experimental results demonstrated that NARX neural network with the Levenberg-Marquardt algorithm for tested architectures, showed a short time in the learning stage. Research limitations/implications: In order to improve the quality of the model in subsequent studies, it would be advisable to introduce additional independent variables as chemical factors. A future research to improve the model may focus on periodic real-time data updates, and to use multiple measurement stations. Social implications: The research will allow local authorities to manage air quality in the city by making decisions in the area of legal regulations and health safety of the local community. The forecasting of PM10 concentration can provides the public with the information with which they can make lifestyle decisions to protect their health. Originality/value: The originality of the article lies in the possibility of using artificial neural networks to forecast the concentration of PM10 from road transport in in urban agglomerations in Poland. Keywords: PM10 forecasting, air quality/pollution monitoring, artificial neural network, road transport, environmental management. Category of the paper: Research paper.