Google Cloud Associate Data Practitioner — Question 74

Your retail company wants to analyze customer reviews to understand sentiment and identify areas for improvement. Your company has a large dataset of customer feedback text stored in BigQuery that includes diverse language patterns, emojis, and slang. You want to build a solution to classify customer sentiment from the feedback text. What should you do?

Answer options

• A. Preprocess the text data in BigQuery using SQL functions. Export the processed data to AutoML Natural Language for model training and deployment.
• B. Develop a custom sentiment analysis model using TensorFlow. Deploy it on a Compute Engine instance.
• C. Use Dataproc to create a Spark cluster, perform text preprocessing using Spark NLP, and build a sentiment analysis model with Spark MLlib.
• D. Export the raw data from BigQuery. Use AutoML Natural Language to train a custom sentiment analysis model.

Correct answer: A

Explanation

Option A is correct because it leverages SQL functions in BigQuery to preprocess the data, ensuring it is clean and ready for training with AutoML Natural Language. Option B, while feasible, requires more manual effort and infrastructure management than necessary. Option C involves using Spark, which can be overly complex for this task, and Option D skips important preprocessing steps that are crucial for accurate sentiment analysis.