The project deals with determining and predicting the type of accident taking place in the city of Austin. The city of Austin has heard of many complaints from cyclists that the city is not doing enough to protect them from motor vehicles.
Thus, in order to deal with this problem and confirm the complaints that have been obtained, the city has complied the data of cyclists which are related to the various incidents that have occurred which would help in reviewing the findings to check if the data could be able to confirm this.
This data would thus help in understanding what are the various possible factors which are leading to the accidents based on the severity of the incident that occurred, and also would help in determining to confirm the complains obtain from the cyclists based on the protection from motor vehicles.
In order to predict the type of accident and confirm the complaints obtained from the cyclists, various machine learning models are implemented which would help to predict the type of accident.
Also based on the various metrices and feature importance it would help to determine which factors need to be considered in order to avoid such accidents.
The Data Analysis Workflow considered here is as follows.
-
Data Preparation
-
Exploratory Data Analysis
-
Data Cleaning
-
Data Visualization
-
Pre Modeling Steps
a. Feature Selection & Extraction
b. Correlation Plot
c. Grouping of categories
d. Label Encoding
-
Model Building
a. Logistic Regression Model
b. Decision Tree Classifier
c. Random Forest Classifier
-
Recommendations & Findings