1. What dataset did you use?

I used a Traffic Crash dataset containing structured crash-level data with the following key attributes:

• Crash time details (Hour, Day of Week, Month)
• Weather conditions
• Road surface condition
• Traffic control device & device condition

The dataset represents historical crash records and includes both environmental and road-safety variables, making it suitable for descriptive analytics and predictive modeling.

2. How did you analyze or prepare the data?

The data was analyzed on focusing on the key attributes that contributes to the crash. Converted the injury counts into meaningful metrics:

• Fatal Rate %
• Average Injuries per Crash

Built Analysis based on the key attributes that gives us the overall picture of which Climatic condition, road type or the Week of the day contributed to more number of crashes.

Leveraged the use of OAC AutoML for the prediction of the Injury count based on the Injury type. Used relevant predictors such as Crash Hour, Weather Condition, Speed Limit, Traffic Control Device impact, Lighting condition for prediction. To handle the class imbalance, I evaluated the class distribution and Used OAC AutoML optimization to improve prediction performance.

3. Who is the intended audience for your visualization?

The intended audience includes: Traffic Safety Authorities, Urban Planning Departments, Emergency Response Teams.

The dashboard is designed for decision-makers who need to understand:

• When crashes are most severe
• What environmental or road factors increase risk
• Where preventive measures should be implemented

4. What is your visualization about, and what question does it address?

The core problem that was focused was : What factors contribute most to severe and fatal crashes, and how can risk be predicted and prevented?

The below questions are addressed in my dashboard:

1. What is the peak crash time?
2. Does higher speed limit increase fatality risk?
3. Are crashes more severe at intersections without traffic control?
4. How does weather influence crash severity?
5. Can we predict crash severity using historical data?

The visualization combines:

• Descriptive analytics (trends, severity distribution)
• Diagnostic analytics (cause analysis)
• Predictive analytics (AI-based severity prediction)

5. Did you use any Oracle Analytics AI features?

Yes.

I used AutoML in Oracle Analytics Cloud to build a classification model and auto insights from the exploratory analysis.