Exam Professional Data Engineer topic 1 question 65 discussion

real-valued can not be null N/A or empty, have to be “0”, so it has to be B.

Should be B

Usually, None values are converted to 0 in data cleaning and preparation process. The key point here is, we don't require any other tool than DataPrep to identify and modify the value

Cloud Dataflow is ideal for scalable data processing and allows for real-time transformations. Logistic regression requires numerical (real-valued) inputs, and null values cannot remain as they are.

Option D: Using null value conversion to 0 is the most correct practice for this case. Accompanying it with a script allows us to implement the necessary logic to handle null cases properly, adapting to the model while maintaining data integrity.

B. Dataprep has the feature to convert it into 0.

0 is still a value, which can add bias in the model and the model will take that into account while making predictions so 'none'

Why not D? keyword is Monitor, B would replace all empty fields and also cause unintended bias.

The questions says "You need to prepare data using a casual method ", thats dataprep and values should be 0 so the right answer is B

No brainer. We need a real value and Dataprep is made for this. Dataflow is mainly for pre-processing before BigQuery ingests the data.

Dataprep is made for this kind of stuff, no reason to use a streaming service such as Dataflow.

gpt:Cloud Dataprep is a data preparation service that can be used to transform, clean and shape data in a visually interactive way. It provides an easy-to-use interface to find and replace null values. Cloud Dataflow is a fully-managed service for executing data processing pipelines, which allows for parallel execution of data processing tasks. However, it requires more expertise to set up and operate than Cloud Dataprep, and is usually used for more complex data processing needs. Therefore, option B is the most suitable approach for the given requirements.

Seems to me like Answers are both B and D. B is faster to implement while D takes time. Doesnt mean that it's wrong though. I m not sure why everyone has picked just B. Why not D? D works and does the same job. And also having custom script provides more flexibility and control over the data processing tasks and it allows you to handle missing values in a more flexible and efficient way.

B is the correct answer.

Answer is Use Cloud Dataprep to find null values in sample source data. Convert all nulls to 0 using a Cloud Dataprep job. Key phrases are "casual method", "need to replace null with real values", "logistic regression". Logistic regression works on numbers. Null need to be replaced with a number. And Cloud dataprep is best casual tool out of given options.

real value 0

It is B