In recent years, traffic issues in China have been emerging continuously, and the
traffic congestion problem in Beijing is particularly prominent. We have
explored the relationships between factors such as driving duration, road
length, weather conditions in Beijing and traffic congestion. By using the
Logistic Regression Model to analyze the relationships among driving duration,
road length and traffic congestion, we found that both driving duration and road
length are negatively correlated with traffic congestion. The model shows high
accuracy and recall rate, demonstrating excellent performance. We also employed
the Weighted Average Correlation Model to study the relationship between weather
conditions and traffic congestion. The results indicate that traffic congestion
is more severe in rain, snow, and foggy weather, while it is less serious in
sunny and cloudy weather. Subsequently, through the noise level verification,
the stability of the model was confirmed. At the same time, we used Shapley
value analysis, Bootstrap confidence intervals, and hypothesis testing to
examine the impacts of travel time and road length on traffic congestion.
Additionally, we employed Cross-validation and Granger causality test to assess
the influence of weather conditions on traffic congestion. The results of these
analyses all verify the correctness of our conclusions. Finally, based on these
results, we put forward suggestions regarding travel arrangements and the
setting of traffic facilities. We Suggests guiding the public to rationally
choose travel modes based on congestion and weather. Points out that logistic
regression and weighted average models have limitations in capturing non-linear
relationships and are sensitive to outliers.