ANL303: Globally, Numerous People Died in Road Traffic Crashes Every Year For the Betterment of the Public at large, Ways to Alleviate the Frequency and Severity of Traffic: Fundamentals of Data Mining Assignment, SUSS, Singapore

Question 1

Globally, numerous people died in road traffic crashes every year. For the betterment of the public at large, ways to alleviate the frequency and severity of traffic crashes have been the primary concerns of many governments. In fact, many traffic crashes can be avoided with the implementation of effective policies and regulations.

(a) Traffic crash analysis can be used to analyse crash data with an objective to reduce crash fatalities and improve road safety.

(i) Give one (1) example of how descriptive data mining can be used to achieve the stated objective.

(ii) Give one (1) example of how predictive data mining can be used to achieve the stated objective.

(b) Assume that a dataset is collected for traffic crash analysis. The dataset contains data on severe traffic crashes between 2015 and 2019 in Australia. The description of the dataset is listed in Table 1.

Suggest two (2) additional variables that can be included in traffic crash analysis. Explain the rationale of their inclusion.

(c) Assume that there are two interesting findings obtained from traffic crash analysis:

• An experienced driver was likely to induce a severe crash on highways even though there were adequate safety facilities.
• The majority of severe crashes occurring in the southwest Australia region were single-vehicle crashes rolling down from a hill.

Give one (1) potential explanation of each finding. Then, propose one (1) solution for improving road safety based on each finding.

(d) Assume that your proposed solutions stated in part (c) have been implemented. Discuss how you are going to evaluate the effectiveness of the solutions.

(e) Identify one (1) limitation of traffic crash analysis using the dataset stated in Table 1 and propose one (1) way to address the limitation. (Note: The limitation should not be about the lack of variables as stated in part (b).)