Using CLAIRE GPT > Using CLAIRE GPT > Data Governance and Catalog
  

Data Governance and Catalog

The following table lists the Data Governance and Catalog tasks that you can perform in CLAIRE GPT:
Task
Description
Data discovery
Harnesses the data discovery capability offered by Data Governance and Catalog and uses the catalog of enterprise assets to quickly search and discover assets. Start conversations to search for data assets relevant to your business use case and discover semantically similar assets. You can find assets by catalog source type, created date, owner, name, and relationships.
For more information about data discovery, see Discover assets from the catalog.
Metadata exploration
Utilizes the metadata exploration capability offered by Data Governance and Catalog to explore, view, and understand the metadata of an asset present in the catalog of enterprise assets. This includes an asset's overview, stakeholders, data profile, data quality score, ratings, lineage, and system attributes.
For more information about metadata exploration, see Explore metadata of assets from the catalog.
Data exploration
Leverages the data exploration capability offered by Data Governance and Catalog to understand the quality and shape of your source data and identify data anomalies.
For more information about data exploration, see Explore source data of assets from the catalog.
Data transformation
Determines the data sets that best meet your business requirements from the catalog and utilizes the data transformation capability of Data Integration to load, transform, and integrate your data. Create ELT pipelines and run them as mappings in Data Integration to transform your data.
For more information about data transformation, see Data transformation.

Fetch data to create a sales performance report

You’re a business user and you’re asked to create a report about sales performance in your organization. With the help of this report, you can identify patterns and forecast sales for the upcoming year.
The sales data is in a Snowflake data warehouse, and your organization cataloged the data in Data Governance and Catalog. To create the report, you need to look for relevant data sets in the catalog for overall sales, online sales, and store sales.
The following image shows the steps you can perform to create a sales performance report:
Steps you can perform to create a sales performance report.
Let’s discover the relevant data sets in Data Governance and Catalog. Then, we’ll view the metadata to decide which data sets to use for the sales performance report.
  1. 1Log in to CLAIRE GPT.
  2. 2Start your conversation to find all the data sets related to sales.
    3. Enter the following prompt:
    Find out all the sales data sets
    The following image shows the response of CLAIRE GPT to the prompt, displaying data sets with names that contain ‘sales’ across all catalog sources:
    Response from CLAIRE GPT displaying data sets with names that contain ‘sales’ across all catalog sources.
    CLAIRE GPT provides a summary of the data sets to help you find the right data.
    This showcases the data discovery capability of CLAIRE GPT.
    For more information about data discovery, see Data discovery.
  3. 3To narrow the results further, search for the ‘sales’ tables that are present only in the Snowflake catalog source type.
    4. Enter the following prompt:
    Show only the sales tables in the Snowflake catalog source type
    The following image shows the response of CLAIRE GPT to the prompt, displaying a list of tables in Snowflake source with names that contain ‘sales':
    Response from CLAIRE GPT displaying a list of tables in a Snowflake source with names that contain ‘sales'.
  4. 4From the list of tables, further explore and view more details about the FACT_SALES table.
    5. Enter the following prompt:
    Show me the overview of @
    When you type @ in the prompt text box, a list of assets that were recently discovered based on your conversation history appears. If you continue to type the asset name, the list shortens, and you can select an asset from this list.
    Select the ‘FACT_SALES’ column.
    The following image shows the response of CLAIRE GPT to the prompt, displaying an overview of the requested FACT_SALES table as shown in the following image:
    Response from CLAIRE GPT displaying an overview of the FACT_SALES table.
    The overview includes a description, data characteristics such as the number of columns, data quality scores, profiling results, key columns, and so on. You can see that the data has a high data quality score, it's been profiled, and it has the key metrics you need for the analysis. You can also see who the stakeholders are so that you can request access to the data.
    This showcases the metadata exploration capability of CLAIRE GPT.
    For more information about metadata exploration, see Metadata exploration.
  5. 5Now, let's find the online sales data in your organization. For this, let's explore the metadata of the FACT_ONLINE_SALES table.
    6. In the prompt, use @ to see suggestions for the names of data sets and select FACT_ ONLINE_SALES or directly enter the prompt.
    Enter the following prompt:
    Show me the columns of FACT_ ONLINE_SALES
    The following image shows the response of CLAIRE GPT to the prompt, displaying columns of the table:
    Response from CLAIRE GPT displaying the columns of the FACT_ ONLINE_SALES table.
    CLAIRE GPT also provides suggestions to help you find the right data.
  6. 6Before you proceed further, you'd want to find out if you’re familiar with the stakeholders of the FACT_ONLINE_SALES table.
    7. Enter the following prompt:
    Who is the stakeholder of FACT_ONLINE_SALES?
    The following image shows the response of CLAIRE GPT to the prompt, displaying the stakeholder’s name:
    Response from CLAIRE GPT displaying the stakeholder of the FACT_ONLINE_SALES table.
  7. 7Let's further explore and ask for the lineage to understand the origin and destination of the online sales table.
    8. Enter the following prompt:
    Show me the lineage of FACT_ONLINE_SALES
    The following image shows the response of CLAIRE GPT to the prompt, displaying the lineage of the table:
    Response from CLAIRE GPT displaying the data lineage of the FACT_ONLINE_SALES table.
    Click on the asset name to see the detailed data set level lineage in Data Governance and Catalog.
    By default, the lineage diagram shows five hops. You can expand the lineage by five hops to the right and five hops to the left. To expand the lineage and to view additional hops, click Expand 5 Hops.
    Click the Download icon to save the lineage diagram for future reference.
  8. 8To understand the online sales patterns, you can view the profiling statistics in the tables that you’ve discovered.
    9. Enter the following prompt:
    Show me the data profile of SalesAmount
    The following image shows the response of CLAIRE GPT to the prompt, displaying the profiling statistics of the SalesAmount column:
    Response from CLAIRE GPT displaying the profiling statistics of the SalesAmount column.
    You can see that the column contains both distinct and non-distinct values.
  9. 9In addition to online sales, you want your report to contain information about store sales.
    10. Enter the following prompt:
    Show me the overview of FACT_STORE_SALES
    The following image shows the response of CLAIRE GPT to the prompt, displaying the descriptions, a preview of the columns, and stakeholder information for the table:
    Response from CLAIRE GPT displaying the overview of the FACT_STORE_SALES table.
    The overview includes a description, data characteristics such as the number of columns, data quality scores, profiling results, key columns, and so on. You can see that the data has a high data quality score, it's been profiled, and it has the key metrics you need for the analysis. You can also see who the stakeholders are so that you can request access to the data.
After discovering the sales data sets and exploring the metadata of a few tables, you decide to use the FACT_SALES, FACT_ONLINE SALES, and FACT_STORE_SALES tables in Snowflake to create a sales performance report for your organization.

Analyze customer data to create a marketing campaign

You’re a data analyst at a retail company and you want to analyze customer data and the purchase pattern to design a targeted marketing campaign. You want to look at the customer data in the data catalog and decide which information will be helpful in your analysis and impactful for your campaign.
To achieve this, you have to first discover the available customer data, explore the source and attributes further to identify the right data set, and then perform data analysis on top of the discovered asset. CLAIRE GPT’s natural language interface is just perfect for this use case. You can simply enter your intent as natural language prompts and CLAIRE GPT will answer all your questions.
The retail customer data is in a Snowflake data warehouse, and your organization cataloged the data in Data Governance and Catalog. First, we'll discover the relevant data sets and explore the metadata of the discovered data in Data Governance and Catalog. Then, we'll explore the source sample data to decide which data sets to use to analyze customer data.
The following image shows the steps you can perform to analyze customer data and the purchase pattern to design a targeted marketing campaign:
Steps you can perform to analyze customer data and the purchase pattern to design a marketing campaign.
  1. 1Log in to CLAIRE GPT.
  2. 2Let’s start by finding out the available retail customer data sets in Snowflake, which is your enterprise data warehouse.
    3. Start a conversation to search and discover assets in the data catalog.
    Enter the following prompt:
    Show me the data sets that can be used to analyze Retail Customer information in Snowflake
    The following image shows the response of CLAIRE GPT to the prompt, displaying data sets with names that contain 'retail' and ‘customer’ across Snowflake catalog sources:
    Response from CLAIRE GPT displaying data sets with names that contain 'retail and ‘customer’ across Snowflake catalog sources.
    CLAIRE GPT provides a summary of the data sets to help you find the right data.
    This showcases the data discovery capability of CLAIRE GPT.
    For more information about data discovery, see Data discovery.
  3. 3To narrow the results further, let's explore the metadata and take a closer look at the CUSTOMERS table in Snowflake.
    4. Enter the following prompt:
    Show me an overview of @
    When you type @ in the prompt text box, a list of assets that were recently discovered based on your conversation history appears. If you continue to type the asset name, the list shortens, and you can select an asset from this list.
    Select the ‘CUSTOMERS’ table.
    The following image shows the response of CLAIRE GPT to the prompt, displaying an extensive overview of the table:
    Response from CLAIRE GPT displaying an overview of the CUSTOMERS table.
    The overview includes a description, data characteristics such as the number of columns, data quality scores, profiling results, key columns, and so on. You can see that the data has a high data quality score, it's been profiled, and it has the key metrics you need for the analysis. You can also see who the stakeholders are so that you can request access to the data.
    This showcases the metadata exploration capability of CLAIRE GPT.
    For more information about metadata exploration, see Metadata exploration.
  4. 4Let's move on to data exploration and look at some of the sample data in this table.
    5. In the prompt text box, use @ to see suggestions for the names of data sets and select CUSTOMERS or directly type the prompt.
    Enter the following prompt:
    Show me a data sample of 10 rows from CUSTOMERS
    The following image shows the response of CLAIRE GPT to the prompt, displaying a sample of 10 rows from the CUSTOMERS table:
    Response from CLAIRE GPT displaying a data sample of 10 rows from the CUSTOMERS table.
    CLAIRE GPT also provides suggestions to help you find the right data.
    To see the SQL code used to fetch the sample data, click Explanation.
    To save the sample data in a CSV file for future reference, click the Download icon.
    This showcases the data exploration capability of CLAIRE GPT.
    For more information about data exploration, see Data exploration.
  5. 5You now know that this table contains customer information. Let us find the related tables to get more information about the customer orders.
    6. Enter the following prompt:
    Show me all the other tables related to CUSTOMERS
    The following image shows the response of CLAIRE GPT to the prompt, displaying a list of all the other tables related to CUSTOMERS table:
    Response from CLAIRE GPT displaying all the other tables related to the CUSTOMERS table.
  6. 6Let us start our analysis using all the related tables. Let’s start by exploring customer address and country information.
    7. Enter the following prompt:
    Get full addresses of customers along with their personal details like city, state, and country
    The following image shows the response of CLAIRE GPT to the prompt, displaying the full addresses of customers along with their personal details such as city, state, and country:
    Response from CLAIRE GPT displaying full addresses of customers along with their personal details such as city, state, and country.
    To verify the SQL code used to fetch data from the source, click Explanation.
    The following image shows the SQL code used to fetch the sample source data:
    Response from CLAIRE GPT displaying the SQL code used to fetch the sample source data.
    This showcases the data transformation capability of CLAIRE GPT.
    For more information about data transformation, see Data transformation.
  7. 7Now, let’s extend this further to get information about the customers who placed the orders and the corresponding product information.
    8. Enter the following prompt:
    Who are the customers who placed the orders and what products did they buy?
    The following image shows the response of CLAIRE GPT to the prompt, displaying details of customers who placed the orders and the products they bought:
    Response from CLAIRE GPT displaying the list of customers who placed the orders and the products they bought.
  8. 8Let us refine this to get only the customers from Australia along with the order quantity and the total amount spent on products.
    9. Enter the following prompt:
    Find customers from Australia along with their order quantity and the total amount spent
    The following image shows the response of CLAIRE GPT to the prompt, displaying details of customers from Australia along with their order quantity and the total amount spent:
    Response from CLAIRE GPT displaying details of customers from Australia along with their order quantity and the total amount spent on products.
  9. 9Let’s augment this further with customer address details to use for the marketing campaign.
    10. Enter the following prompt:
    Add emails, addresses, and country information as well
    The following image shows the response of CLAIRE GPT to the prompt, displaying the e-mail, address, city, country, quantity, and total amount spent by customers from Australia:
    Response from CLAIRE GPT displaying email, address, city, country, quantity, and total amount spent by customers from Australia.
    You can copy or download this list and use it for further analysis. You can continue drilling down the table until you get all the details you need for the marketing campaign. This reduces your workload significantly.
We were able to rapidly identify assets of interest and perform metadata and data exploration using simple natural language prompts.