Analysis of Road Accidents
Predict severity of accidents from combinations of road factors.
Task
Is it possible to predict the likely severity of an accident from certain combinations of factors?
Key Requirement: Predict severity of accidents from combinations of road factors.
Delivered: Logistic Regression using R and PowerBI.
Data on UK Road Accidents from at least 2008 contiguously to currently 2016 are publicly available at data.gov.uk/dataset/road-accidents-safety-data and contain. For each year, the minimum 3 files (and sometimes additional files) are Accidents, Casualties and Vehicles. The first 2 of these were analysed in R to potentially answer the following question:
Is it possible to predict the likely severity of an accident from certain combinations of factors?
The aim to determine which factors are common to accidents of different severity levels is streamlined for this report, to focus on which factors have an estimated likelihood to result in non-trivial casualty severity. This would reject the “Null Hypothesis” – that any measurements of relevant factors are purely down to chance.
- Data Preparation and Re-Structuring
- Exploratory Data Analysis
- SQL and R Programming
- Logistic Regression in R
- Deeper Insight Provision