Engineer Data in Google Cloud

Engineer-Data-in-Google-Cloud-Challenge-Lab

Engineer Data in Google Cloud: Challenge Lab 

GSP327

CREATE OR REPLACE TABLE

  taxirides.taxi_training_data AS

SELECT

  (tolls_amount + fare_amount) AS fare_amount,

  pickup_datetime,

  pickup_longitude AS pickuplon,

  pickup_latitude AS pickuplat,

  dropoff_longitude AS dropofflon,

  dropoff_latitude AS dropofflat,

  passenger_count AS passengers,

FROM

  taxirides.historical_taxi_rides_raw

WHERE

  RAND() < 0.001

  AND trip_distance > 0

  AND fare_amount >= 2.5

  AND pickup_longitude > -78

  AND pickup_longitude < -70

  AND dropoff_longitude > -78

  AND dropoff_longitude < -70

  AND pickup_latitude > 37

  AND pickup_latitude < 45

  AND dropoff_latitude > 37

  AND dropoff_latitude < 45

  AND passenger_count > 0

###############################################################################################################################3

CREATE OR REPLACE MODEL taxirides.fare_model

TRANSFORM(

  * EXCEPT(pickup_datetime)

  , ST_Distance(ST_GeogPoint(pickuplon, pickuplat), ST_GeogPoint(dropofflon, dropofflat)) AS euclidean

  , CAST(EXTRACT(DAYOFWEEK FROM pickup_datetime) AS STRING) AS dayofweek

  , CAST(EXTRACT(HOUR FROM pickup_datetime) AS STRING) AS hourofday

)

OPTIONS(input_label_cols=['fare_amount'], model_type='linear_reg') 

AS

SELECT * FROM taxirides.taxi_training_data

####################################################################################################################################

CREATE OR REPLACE TABLE taxirides.2015_fare_amount_predictions

  AS

SELECT * FROM ML.PREDICT(MODEL taxirides.fare_model,(

  SELECT * FROM taxirides.report_prediction_data)

)​