Predicting RTC Severity using Machine Learning

Local Chapter Liverpool, England Chapter

Coordinated by,

Project background.

UK RTCs which have resulted in a persons death have been on a downward trend since the 1960s – however in 2020 1,516 people lost their lives on UK roads. The UK road systems, especially in Liverpool, are dated which means they have not been upgraded to reflect the increase of cars on the road. This means there are still preventative measures that could be implemented to prevent even more deaths on UK roads.

The UK government compiles and disseminates extensive data about road incidents around the nation (often once per year). This data is particularly fascinating and thorough for analysis and research because it contains, but is not limited to, geographic areas, weather conditions, vehicle types, casualty numbers, and vehicle manoeuvres.

The problem.

By harnessing the power of Machine Learning we intend to predict the severity of RTCs and RTC hotspots which would allow the local authority to implement further traffic safety measures.

Project goals.

1. Classifying RTC severity. 2. Identifying areas with the highest number of RTCs. 3. Identifying what type of vehicles are involved in most RTCs. 4. Monitoring the rate of RTCs over time.

Project plan.

  • Week 1

    1. Data preprocessing

  • Week 2

    2. Exploratory Data Analysis to draw insights

  • Week 3

    3. Feature Engineering – creating new features based on insights drawn from EDA.

  • Week 4

    4. Model Development

  • Week 5

    5. Model Evaluation and Deployment – perhaps on AWS or Google Cloud.

Share project on: