SnowPro Advanced Data Scientist Practice Exam

Join Us Among the Stars

Sign Up & unlock 100% of Exam Questions

Log in / Sign up

No Strings Attached!

90Practice Questions

3Study Modes

Free

TopicSort

Question 1

Train and use machine learning models

A Data Scientist wants to get the coefficient of determination for two columns in a table.
Which function can be used for that?

A REGR_INTERCEPT
B REGR_SLOPE
C REGR_R2
D REGR_SXY

Question 2

Use GenAI and LLM capabilities in Snowflake

Question 3

Use GenAI and LLM capabilities in Snowflake

Question 4

Train and use machine learning models

Question 5

Data science concepts

Question 6

Implement Snowflake data science best practices

Question 7

Data science concepts

Question 8

Data science concepts

Question 9

Implement Snowflake data science best practices

Question 10

Data science concepts

Question 11

Train and use machine learning models

Question 12

Data science concepts

Question 13

Train and use machine learning models

Question 14

Train and use machine learning models

Question 15

Train and use machine learning models

Question 16

Data science concepts

Question 17

Prepare data and use feature engineering in Snowflake

Question 18

Prepare data and use feature engineering in Snowflake

Question 19

Data science concepts

Question 20

Train and use machine learning models

Question 21

Data science concepts

Question 22

Train and use machine learning models

Question 23

Use GenAI and LLM capabilities in Snowflake

Question 24

Implement Snowflake data science best practices

Question 25

Implement Snowflake data science best practices

Page 1 of 4 • Questions 1-25 of 90

1 2 3 4

→

Know a question that should be here? Contribute to this exam

A Data Scientist noticed a high usage on Snowflake Cortex LLM functions.
Which command will restrict user access?

A REVOKE ROLE SNOWFLAKE.CORTEX_USERFROM ROLE PUBLIC;
B DROP DATABASE ROLE SNOWFLAKE.CORTEX_USER;
C DROP ROLE SNOWFLAKE.CORTEX_USER;
D REVOKE DATABASE ROLE SNOWFLAKE.CORTEX_USERFROM ROLE PUBLIC;

A Data Scientist is tasked with gathering a summary of each customer review. The data is stored in a Snowflake table in a column named reviews.
How should they do this with the LEAST operational overhead?

A Use the SQL function snow1ake.cortex.summarize (‘reviews’) to summarize using the built-in summarization model.
B Use the SQL function snowlake.cortex.embed('e5-base-v2, ‘reviews’) to summarize using the built-in text embedding model
C Use the SQL function call snowlake.cortex.complete(‘llama3.1-405b’, ‘reviews’) to summarize with the llama3.1-405b model.
D Use Snowpark Container Services to deploy the llama3.1-405b model from the Hugging Face Hub and submit the ‘text_data’ column to the model API.

A Data Scientist trained a neural network with 50 epochs. After training the network, the results showed a high accuracy on training data but a fairly low accuracy on test data (for example, overfitting problem).
Which technique would help in resolving the overfitting issue? (Choose two.)

A Use regularization.
B Use gradient descent
C Modify the learning rate.
D Implement early stopping,
E Add additional features to the training data.

A Data Scientist has noticed that a model is producing more accurate predictions.
What metric should be used to evaluate why the model predictions have improved?

A Confidence intervals
B Shapley Additive exPlanations (SHAP)
C Root Mean Squared Error (RMSE)
D R2

Question 6

Implement Snowflake data science best practices

Question 7

Data science concepts

Question 8

Data science concepts

Question 9

Implement Snowflake data science best practices

Question 10

Data science concepts

Question 11

Train and use machine learning models

Question 12

Data science concepts

Question 13

Train and use machine learning models

Question 14

Train and use machine learning models

Question 15

Train and use machine learning models

Question 16

Data science concepts

Question 17

Prepare data and use feature engineering in Snowflake

Question 18

Prepare data and use feature engineering in Snowflake

Question 19

Data science concepts

Question 20

Train and use machine learning models

Question 21

Data science concepts

Question 22

Train and use machine learning models

Question 23

Use GenAI and LLM capabilities in Snowflake

Question 24

Implement Snowflake data science best practices

Question 25

Implement Snowflake data science best practices

A Data Scientist executed this command to create a task:

Question Image

What is the reason?

A The task has not been set to resume.
B The cron job must be specified in UTC.
C The SQL is invalid for calling a stored procedure.
D The ALLOW_OV ERLAPPING_EXECUTION command is not valid with a fixed SCHEDULE.

A Data Scientist wants to capture customer profiles likely to churn for a company. The training data set has 2% churn and 98% no churn, and they set 1 to churn and 0 to no churn. Then the Data Scientist ran the model in Snowpark.
After capturing the confusion matrix, what value could be misleading on how the model is performing due to an imbalanced data set?

A Accuracy
B F-measure
C Precision
D Recall

Which algorithms are examples of supervised learning? (Choose two.)

A Apriori
B CART
C K-Means
D PCA
E SVM

A Data Scientist created a session with Snowpark Python (snowflake-snowpark-python package) with this code:

Question Image

Which parameters are required to authenticate the session? (Choose two.)

A Account
B Database
C Role
D User
E Virtual warehouse

What does variance measure in statistics?

A The variability from the minimum value
B The variability from the median value
C The variability from the mean value
D The variability from the mode value

While logging a model with the Snowflake Model Registry by calling the registry’s log_model method, which arguments are required? (Choose two.)

A code_paths
B model
C model_name
D python_version
E version_name

A Data Scientist needs to present results from a linear regression model used to estimate the quality of a user’s experience. The model performed measurements using multiple standardized numeric features that were then applied to a numeric satisfaction scale.
What diagram should be used to present the effect of each feature on user satisfaction scores?

A Confusion matrix
B Correlation matrix
C Scatterplot
D Coefficient plot

A Data Scientist has a trained a model that has been stored as a file on a Snowflake stage. Now they want to use a Python User-Defined Function (UDF) for scoring data in Snowflake with the model
One requirement is that they should be able to replace the mode file with a new one without having to recreate the UDF.
What is required in order to accomplish this?

A Load the model file using the path to the model file including stage and file name in the UDF.
B Deploy the UDF using Snowpark for Python and use the trained model object in the UDF.
C Provide the path to the model file including stage and file name as a parameter to the function and load the file in the UDF.
D Add the path to the model file including stage and file name to the IMPORT parameter for the UDF and load the file in the UDF.

A Data Scientist at a telecommunications company wants to store churn scores that are calculated with a Snowflake external function, configured to connect a model endpoint.
The environment conditions are as follows:
The External Function DDL:

Question Image

The CUSTOMER_DAILY table:

Question Image

The churn value stored in a new table called CUSTOMER_CHURN:

Question Image

Based on the environmental conditions, how could the Data Scientist persistently store the churn scores from the external function in Showflake?

A insert into customer_churnselectcust_id,get_churn(*) as churn_scorefromcustomer_daily;
B insert into customer_churnselectcust_id,call get_churn(object_construct(*)) as churn_scorefromcustomer_daily;
C insert into customer_churnselectcust_id,call get_churn(*) as churn_scorefromcustomer_daily;
D insert into customer_churnselectcust_id,get_churn(object_construct(*)) as churn_scorefromcustomer_daily;

Which features can be used with the Snowflake Model Registry to seamlessly manage models from development to production lifecycles? (Choose two.)

A Tags
B Zero-copy clones
C Account replication
D Model Copy
E Time Travel

A Data Scientist is training a model. They used scikit-learn to create a partial dependency plot using the following command:

Question Image

The following charts are created as a result:

Question Image

Which statements reflect what is depicted in the chats? (Choose three.)

A As Median income increases, the target housing price decreases.
B As Median income increases, the target housing price increases.
C As the Age of the House increases, the target housing price decreases.
D As the Age of the House increases, the target housing price is fairly constant.
E As the Average Occupancy increases, the target housing price decreases.
F As the Average Occupancy increases, the target housing price increases.

This correlation matrix was created when performing feature engineering:

Question Image

Which combination of variables is the MOST correlated and could possibly help with feature reduction?

A DIS and CRIM
B DIS and TAX
C RAD and TAX
D INDUS and CRIM

A Data Scientist is building a data pipeline for a customer churn model. To enable efficient processing of the model, they add a stream to the customer table.
Which function should be used to check if the stream has new or updated data?

A SYSTEM$STREAM_STATUS(‘MYSTREAM’)
B SYSTEM$STREAM_HAS_DATA(‘MYSTREAM')
C SYSTEM$STREAM_GET UPDATES (‘MYSTREAM’)
D SYSTEM$STREAM_GET_ TABLE_TIMESTAMP (‘MYSTREAM’)

A Data Scientist executes a SQL NULL argument to a Python User-Defined Function (UDF) in a Snowflake string data type.
What will be returned as a translated Python value?

A 1
B False
C None
D An empty string

Which step of the machine learning lifecycle does hyperparameter tuning fall under?

A Model training
B Model deployment
C Model validation
D Feature engineering

A remote weather sensor malfunctions and produces temperature readings higher than the normal range which was around 69.8°F (21°C).
Ignoring units, what is the correct order of the magnitude of these key measures?

A Mean> Median > Skewness > 0(A right skew and a mean greater than the median)
B Median > Mean > Skewness > 0(A right skew and a median greater than the mean)
C Mean > Median > 0 > Skewness(A left skew and a mean greater than the median)
D Median > Mean > 0 > Skewness(A left skew and a median greater than the mean)

This chart in Snowsight for a New York City ride-share bike service shows the number of trips taken to the destination borough:

Question Image

A Data Scientist wants to build a classifier that predicts which borough will be the most likely destination when a trip is initiated.
Which techniques should be used to handle the class imbalance depicted in the Snowsight chart? (Choose two.)

A Introduce regularization parameters.
B Collapse all minority classes into a single class.
C Undersample the Manhattan borough for training.
D Utilize bootstrapping to synthesize additional data for the non-Manhattan boroughs.
E Utilize Synthetic Minority Oversampling Technique (SMOTE) to oversample the non-Manhattan boroughs for training.

For which Snowflake Cortex LLM functions would both input and output tokens be counted? (Choose three.)

A SUMMARIZE
B SENTIMENT
C TRANSLATE
D EMBED_TEXT_768
E EMBED_TEXT_1024
F CLASSIFY_TEXT

A company’s platform team wants to integrate their existing data lake with Snowflake. The data lake is hundreds of TBs in size and the team does not want to duplicate most of the data into Snowflake. A Data Scientist at the company wants to be able to query and access the data lake’s metadata. There is already an external stage in Snowflake referencing the data lake’s location.
What is the MOST efficient way to integrate the existing data lake into the Snowflake environment?

A Move the data lake files to an internal stage in Snowflake to allow for access to the metadata and data from the data lake.
B Using the existing external stage, create SELECT statements that can be run on-demand, referencing the files in the data lake directly.
C Create a PIPE for each data lake table which will allow for on-demand querying of the data lake files by leveraging the Snowpipe service.
D Create an external table for each corresponding data lake table to enable querying data stored in files in the data lake as if the data lake table was inside a database.

Which characteristic applies to Snowpark Python stored procedures?

A They must return a value.
B They need to have session as an input.
C They have to contain a process method
D They cannot contain Snowpark API calls.

Join Us Among the Stars

SnowPro Advanced Data ScientistPreview

About the SnowPro Advanced Data Scientist Exam

Mode Selection

Question 1

Question 2

Question 3

Question 4

Question 5

Question 6

Question 7

Question 8

Question 9

Question 10

Question 11

Question 12

Question 13

Question 14

Question 15

Question 16

Question 17

Question 18

Question 19

Question 20

Question 21

Question 22

Question 23

Question 24

Question 25

Question 6

Question 7

Question 8

Question 9

Question 10

Question 11

Question 12

Question 13

Question 14

Question 15

Question 16

Question 17

Question 18

Question 19

Question 20

Question 21

Question 22

Question 23

Question 24

Question 25