Product Engineering Review: Remaining_Useful_Life_Forecasting_PY_SQL

[chetan-hirapara]

Reviewer 1 comments:

• Use teradataml API’s as much as possible instead of queries.
• to_pandas() pull the data to client. Do not use it unless it is really required.

Reviewer 1 suggestions:

• use df.group_by instead of query and use avg and count functions.
• teradataml supports function corr use the same instead of pandas.corr.
• use teradataml TDDecisionForest, TDDecisionForestPredict function instead of query.
• teradataml supports in-house visualisation. Use teradataml DataFrame.plot instead of pandas/matplotlib plotting.

Reviewer 2 comments:

• use of sql queries can be avoided at many places.

Reviewer 1 suggestions:

• we have evaluate method instead of RegressionEvaluator in latest teradataml versions we can use that.