Data Sources

Assume that you have the following data sources available to you.

Intersection

Variable	Data Type	Example
Location ID	UUID	1XAA-BDDS-XX41
Signal	Boolean	Yes
Number of Lanes	Numeric	4
Pedestrian Crossing	Boolean	Yes

Road Segment

Variable	Data Type	Example
Location ID	UUID	3430-33XD-4443
Intersection A ID	UUID	1XAA-BDDS-XX41
Intersection B ID	UUID	1XAA-4FDK-ZX41
Number of Lands	Numeric	Yes
One Way	Boolean	No
Segment Type	Categorical	(Highway, Road, Dirt)
Length (Feet)	Numeric	50
Speed Limit (MPH)	Numeric	70

Driver Information

Variable	Data Type	Example
Driver ID	UUID	34GA
Vehicle Type	Categorical	(SUV, Sedan)
Ratings	Numeric	4.9
Age	Numeric	68

Historical Trip

Variable	Data Type	Example
Trip ID	UUID	566FJA-X4315
Start ID	UUID	433D-55XD-7743
End ID	UUID	3430-33XD-4443
Start Time	Timestamp	July 28, 2023 15:33:25.15
End Time	Timestamp	July 28, 2023 15:53:25.15
Distance (Mile)	Numeric	5.3
Time (Minutes)	Numeric	20

Weather Data

Variable	Data Type	Example
Location ID	UUID	3430-33XD-4443
Timestamp	Timestamp	July 28, 2023 15:53:25.15
Condition	Categorical	(Sunny, Rain, Foggy)

Feature Engineering

Intersection ETA Model Features – Features that are useful in predicting the average waiting time at an intersection

Features	Data Type	Data Source	Feature Engineering
Signal	Boolean	Intersection	One Hot Encoding
Number of Lanes	Numeric	Intersection	Mean Imputation
Pedestrian Crossing	Boolean	Intersection	One Hot Encoding
Vehicle Type	Categorical	Driver	One Hot Encoding
Ratings	Numeric	Driver	Mean Imputation
Age	Numeric	Driver	Mean Imputation
Start Time	Time	Historical Trip	Time Decomposition (year, month, day, hour,…)
Condition	Categorical	Weather	One Hot Encoding

Road Segment ETA Model Features – Features that are useful in predicting the average traveling time on a road segment from one intersection to the next.

Features	Data Type	Data Source	Feature Engineering
Number of Lands	Numeric	Road Segment	Mean Imputation
One Way	Boolean	Road Segment	One-Hot Encoding
Segment Type	Categorical	Road Segment	One-Hot Encoding
Length (Feet)	Numeric	Road Segment	Mean Imputation
Speed Limit (MPH)	Numeric	Road Segment	Mean Imputation
Vehicle Type	Categorical	Driver	One Hot Encoding
Ratings	Numeric	Driver	Mean Imputation
Age	Numeric	Driver	Mean Imputation
Start Time	Time	Historical Trip	Time Decomposition (year, month, day, hour,…)
Condition	Categorical	Weather	One Hot Encoding

Feature Selection

You can consider filter, wrapper or model-based feature selection that utilizes L1, L2, or Elastic-Net regularization. Dimensionality reduction using PCA could be considered too.

Data Preparation

Unlock now by joining.