Spark ML · Churn Intelligence

Accuracy—

Precision—

Recall—

F1 score—

ROC AUC—

Customers flagged—

ROC performance

threshold independent

Confusion matrix

live threshold

Risk score distribution

stayedchurned

Learned feature influence

model coefficients

Decision landscape

usage drop vs probability

Highest-risk customers

batch predictions

Customer	Risk	Action	Observed	Strongest signal

Equivalent PySpark pipeline

copy-ready

# The browser calls the same spark-ml estimator compiled to WASM
from pyspark.ml import Pipeline
from pyspark.ml.feature import VectorAssembler, StandardScaler
from pyspark.ml.classification import LogisticRegression

assembler = VectorAssembler(inputCols=["tenure", "monthly_spend", "support_calls", "usage_drop", "late_payments"], outputCol="features")
scaler = StandardScaler(inputCol="features", outputCol="scaled", withMean=True, withStd=True)
lr = LogisticRegression(featuresCol="scaled", labelCol="churned", maxIter=800, threshold=0.50)
model = Pipeline(stages=[assembler, scaler, lr]).fit(train_df)
predictions = model.transform(customer_batch)
predictions.select("customer_id", "probability", "prediction").show()

Algorithm spark_ml::LogisticRegressionTraining rows —WASM fit time —