Bike Rental Prediction

This workflow reads in a dataset.It then Predicts the number of bikes to be rented in any given hour.

Workflow

Below is the workflow. It does the following:

  • Reads data from a sample dataset.
  • Extracts hour from time using datatype timestamp.
  • Calculates Count to datatype double.
  • Assembles features for modelling.
  • Calculates vectorindexer.
  • Splits it.
  • GBTRegression.
  • Prediction.
  • RegressionEvaluator.
  • Correlation with columns.
  • Summary analysis.
  • Calculate count for rental per hour.
  • Analyse using Graph.
Bike Rental Prediction

Reading from Dataset

It reads sample Dataset file.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Extract hour from time using datatype timestamp

It Extracts hour from time using datatype timestamp using DateTimeFieldExtract Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Calculate Count to datatype double

It Calculates cast the Count field to datatype double using CastColumnType Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Assemble features for modelling

It Assembles features columns into a feature vector using VectorAssembler Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Calculate vectorindexer

It identifies categorical features and index them using vectorindexer Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Split it

It will split our dataset into seperate training and test sets using split Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

GBTRegression

It validates held out test sets in order to know about high confidence using GBTRegression Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Prediction

It will make prediction on future data using Prediction Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

RegressionEvaluator

It validates held out test sets in order to know about high confidence using RegressionEvaluator Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Correlation with columns

It will analyse correlation between various columns using Correlation Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Summary analysis

It visualizes our data to get sense of whether the features are meaningful using Summary Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Calculate count for rental per hour

It calculates count for rental per hour using query with SQL Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction

Analyse using Graph

It will analyse graph with bike rental counts and hours of the day using GraphValue Node.

Processor Configuration

Bike Rental Prediction

Processor Output

Bike Rental Prediction