https://github.com/ravi72munde/scala-spark-cab-rides-predictions

A big data project for predicting prices of Uber/Lyft rides depending on the weather
https://github.com/ravi72munde/scala-spark-cab-rides-predictions

predict-prices scala spark spark-streaming streaming uber weather

Last synced: 24 days ago
JSON representation

A big data project for predicting prices of Uber/Lyft rides depending on the weather

• Host: GitHub
• URL: https://github.com/ravi72munde/scala-spark-cab-rides-predictions
• Owner: ravi72munde
• License: mit
• Created: 2018-11-19T06:17:22.000Z (about 6 years ago)
• Default Branch: master
• Last Pushed: 2024-03-19T17:57:15.000Z (10 months ago)
• Last Synced: 2024-03-19T19:01:11.185Z (10 months ago)
• Topics: predict-prices, scala, spark, spark-streaming, streaming, uber, weather
• Language: Scala
• Size: 1.58 MB
• Stars: 12
• Watchers: 3
• Forks: 8
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# scala-spark-cab-rides-predictions

A big data project for predicting prices of Uber/Lyft rides depending on the weather.

Dataset was compiled and uploaded to Kaggle. Can be found here https://www.kaggle.com/ravi72munde/uber-lyft-cab-prices

## Contributors:

* Ravi Munde

* Karan Barai

### Project Structure :

* cab-price-connector - Data Collection Scala Project

* Databricks_Prediction_code.html - Anlysis and Spark Model(From Databricks.com)

* Cab_Price_Prediction.ipynb - Random Forrest Model in Python

### Data Model:

#### CabPrice

root

     |- cab_type : String

      |- destination : String

      |- distance: Float

      |- id: String

      |- name: String

      |- price: Float

      |- product_id: String

      |- source: String

      |- surge_multiplier: String

      |- time_stamp:Long

#### Weather

root

    |- clouds : Float

    |- humidity : Float

    |- location : Float

    |- location : String

    |- temp : String

    |- pressure : Float

    |- wind : Float

![Actor System](Actors.png)

Sample log of Actor System Running on EC2

`INFO [CabRideSystem-akka.actor.default-dispatcher-2] a.DynamoActor - received 12 number of weather records`

`INFO [CabRideSystem-akka.actor.default-dispatcher-4] a.DynamoActor - Weather Batch processed on DynamoDB`

`INFO [CabRideSystem-akka.actor.default-dispatcher-9] a.DynamoActor - received 156 number of cab price records`

`INFO [CabRideSystem-akka.actor.default-dispatcher-8] a.DynamoActor - Cab Prices Batch processed on DynamoDB`

`INFO [CabRideSystem-akka.actor.default-dispatcher-7] a.Master - Cab ride data piped to Dynamo Actor`

`INFO [CabRideSystem-akka.actor.default-dispatcher-13] a.DynamoActor - received 156 number of cab price records`

`INFO [CabRideSystem-akka.actor.default-dispatcher-15] a.DynamoActor - Cab Prices Batch processed on DynamoDB`

*NOTE: AWS Creditials need to be put in environment vairables*

### Model Evaluation Matrices

* Regression R_squared = 0.62

* Random Forrest Regression's Price Prediction Accuracy : 92.79 %

* Random Forrest Classification Surge Prediction Accuracy: 77.69 %

  

Confusion Matrix for the Classifier