Update the python notebooks for the customer churn examples

[NvTimLiu]

1, Add the variables 'SPARK_MASTER_URL' and 'DATA_ROOT' to support automated testing from CI/CD jobs.

2, 'output_prefix' is NOT referenced in below lines

[NvTimLiu]

Build PASS:

[YanxuanLiu]

Does it still require same version for spark cluster and pyspark?

[NvTimLiu]

Does it still require same version for spark cluster and pyspark?

Yes we need to keep the pyspark and spark-bin version the same,

actually we only need to point the PYTHONPATH to the current spark binary [export PYTHONPATH=$SPARK_HOME/python].

[YanxuanLiu]

Does it still require same version for spark cluster and pyspark?

Yes we need to keep the pyspark and spark-bin consistent,

actually we only need to point the PYTHONPATH to the current spark binary [export PYTHONPATH=$SPARK_HOME/python].

Can we avoid spark session not found error by pointing PYTHONPATH only but ignoring the pyspark version installed?

[NvTimLiu]

Does it still require same version for spark cluster and pyspark?

Yes we need to keep the pyspark and spark-bin consistent,
actually we only need to point the PYTHONPATH to the current spark binary [export PYTHONPATH=$SPARK_HOME/python].

Can we avoid spark session not found error by pointing PYTHONPATH only but ignoring the pyspark version installed?

yes, right

[YanxuanLiu]

LGTM. We also need to update the README to introduce the environment we need to execute the notebook.

[nvliyuan]

LGTM