Using CLAIRE GPT > Data Governance and Catalog > Analyze customer data to create a marketing campaign
  

Analyze customer data to create a marketing campaign

You’re a data analyst at a retail company and your organization's retail customer data is stored in a Snowflake data warehouse, and cataloged in Data Governance and Catalog. You want to analyze customer data and the purchase pattern to design a targeted marketing campaign. You want to look at the details associated with customer purchase patterns in the data catalog and decide which information will be helpful in your analysis and impactful for your campaign.
To achieve this goal, you can take the following approach:
Following the approach, you can analyze the purchase pattern of customers and design a targeted marketing campaign.
For more information about the discovery agent, see Discovery agent.
To achieve the goal, perform the following steps:
  1. 1Log in to CLAIRE GPT.
  2. 2Start a conversation to search customer data sets in Snowflake catalog sources that are profiled, have high data quality scores, and contain critical data elements.
    3. Enter the following prompt:
    I want to analyze customer data for a marketing campaign. Show me the data sets from Snowflake that have good data quality, are profiled, and have critical data elements.
    The following image shows the response displaying data sets across Snowflake catalog sources that are profiled, have high data quality scores, and contain critical data elements:
    The response displays multiple customer data sets, such as CUSTOMER_ANALYTICS, CUSTOMERSPROFILE, and CUSTOMER. It also displays a card named Customer datasets from Snowflake with good data quality, profiling, and critical data elements.
    To view the data sets, click the card named Customer datasets from Snowflake with good data quality, profiling, and critical data elements.
    You can expand the reasoning to view the analysis process of the discovery agent.
    The following image shows the canvas with a tabular view of the customer data sets in Snowflake:
    The canvas displays the two customer data sets, CUSTOMER_ANALYTICS and CUSTOMERSPROFILE, in a tabular view. The right side panel contains contextual information about the data.
    You can click an asset name to open it in Data Governance and Catalog.
  3. 3To understand more about a data set, you can explore the metadata and take a closer look at the CUSTOMERSPROFILE table in Snowflake.
    4. You can either type the number corresponding to the data set or type the data set name, CUSTOMERSPROFILE.
    The following image shows the response displaying an extensive overview of the CUSTOMERSPROFILE table:
    The response displays an overview of the CUSTOMERSPROFILE table that includes data characteristics such as data quality results, key columns, relationships, hierarchy details, and so on. It also displays a card named CUSTOMERSPROFILE Asset Overview.
    You can see that the data has a high data quality score, it has been profiled, and it has the key metrics you need for the analysis. You can also see who the stakeholders are in case you need to request access to the data.
    To view the overview in detail, click the card named CUSTOMERSPROFILE Asset Overview.
    The following image shows the canvas with a detailed overview of the CUSTOMERSPROFILE table:
    The canvas displays an overview of the CUSTOMERSPROFILE table with data quality scores and data elements. The right panel displays a summary of the CUSTOMERSPROFILE table.
  4. 4To explore source data, you can look at some of the sample data in this table.
    5. Enter the following prompt:
    Give me a list of customers who have placed orders, including the products they purchased, as well as each customer’s complete address and contact details.
    To get this information, CLAIRE GPT identifies and joins the CUSTOMERSPROFILE table with the ORDERS and PRODUCTS tables.
    The following image shows the response displaying a list of customers who have placed orders along with the products they bought and their full addresses and contact details:
    The response displays a summary of customers who have placed orders along with the products they bought and their full addresses and contact details. It also displays a card named Customer Orders with Product and Contact Details.
    To preview the source data, click the card named Customer Orders with Product and Contact Details.
    The following image shows the canvas with a sample of source data after joining the CUSTOMERSPROFILE table with the ORDERS and PRODUCTS tables:
    The canvas displays a preview of 100 rows of source data after joining the CUSTOMERSPROFILE table with the ORDERS and PRODUCTS tables in a tabular view. The right panel displays a summary, an Explanation, and a Show Code option.
    You can copy or download this list and use it for further analysis. To copy and paste the sample data to a CSV file, click the Copy icon. To export the sample data to a CSV file, click the Download icon.
    Click Explanation for a comprehensive explanation on how the data is fetched. To see the SQL code used to fetch the sample data, click Show Code.
  5. 5To design the marketing campaign, you can analyze trends by age group for the top-selling product.
    6. Enter the following prompt:
    Analyze trends by age group for top product
    To get this information, CLAIRE GPT identifies the top-selling product, segments customers by age group, and analyzes purchase trends for the top-selling product across different age groups.
    The following image shows the response displaying the analysis details:
    The response displays the plan to analyze trends by age group for the top-selling product. It also displays a card named Purchase Count by Age Group for Top-Selling Product.
    To see the analysis details, click the card named Purchase Count by Age Group for Top-Selling Product.
    The following image shows the canvas with a sample of source data in a bar chart:
    The canvas displays a bar chart with the number of purchases in the x axis and age group in the y axis. The right panel displays a summary, an Explanation, and a Show Code option.
    By default, a bar chart appears. You can switch between tables, bar charts, pie charts, and line charts.
You can continue to drill down the data until you get all the details you need for the marketing campaign. This reduces your workload significantly.
You were able to quickly identify assets of interest and perform data discovery, metadata exploration, and data exploration using simple natural language prompts.