Skip to main content
Version: 1.0.1

Model Deployment with Spark Serving

In this example, we try to predict incomes from the Adult Census dataset. Then we will use Spark serving to deploy it as a realtime web service. First, we import needed packages:

Now let's read the data and split it to train and test sets:

data =
data =["education", "marital-status", "hours-per-week", "income"])
train, test = data.randomSplit([0.75, 0.25], seed=123)

TrainClassifier can be used to initialize and fit a model, it wraps SparkML classifiers. You can use help( to view the different parameters.

Note that it implicitly converts the data into the format expected by the algorithm. More specifically it: tokenizes, hashes strings, one-hot encodes categorical variables, assembles the features into a vector etc. The parameter numFeatures controls the number of hashed features.

from import TrainClassifier
from import LogisticRegression

model = TrainClassifier(
model=LogisticRegression(), labelCol="income", numFeatures=256

After the model is trained, we score it against the test dataset and view metrics.

from import ComputeModelStatistics, TrainedClassifierModel

prediction = model.transform(test)
metrics = ComputeModelStatistics().transform(prediction)

First, we will define the webservice input/output. For more information, you can visit the documentation for Spark Serving

from pyspark.sql.types import *
from import *
import uuid

serving_inputs = (
.address("localhost", 8898, "my_api")
.option("name", "my_api")
.parseRequest("my_api", test.schema)

serving_outputs = model.transform(serving_inputs).makeReply("prediction")

server = (
.option("checkpointLocation", "file:///tmp/checkpoints-{}".format(uuid.uuid1()))

Test the webservice

import requests

data = '{"education":" 10th","marital-status":"Divorced","hours-per-week":40.0}'
r =, url="http://localhost:8898/my_api")
print("Response {}".format(r.text))
import requests

data = '{"education":" Masters","marital-status":"Married-civ-spouse","hours-per-week":40.0}'
r =, url="http://localhost:8898/my_api")
print("Response {}".format(r.text))
import time

time.sleep(20) # wait for server to finish setting up (just to be safe)