Recipes

Recipes are quickstarts for popular predictions, combining a use case with its connection to create a fully-provisioned pipeline.

Recipes are your quickstart guides for popular use cases with Faraday. Browse through dozens of popular use cases in Faraday's recipe library, from assigning personas in BigQuery, where you’re free to deliver them to your favorite activation platform, to scoring leads in Salesforce. The entire rollout process is presented for you from start to finish in one easy screen. Regardless of what you have configured in your Faraday account–maybe you have a connection and a dataset, but no cohorts yet (or maybe nothing at all!)–the wizard will guide you through just what you need to build your chosen recipe. From a cold start to predictions in your stack, in just a few clicks.

Getting started

  1. Select a recipe from the list.

📘Recommended recipes

  1. In the recipe wizard, you'll be presented with the required parts of the recipe, which you'll be able to create (if necessary) or select below. In this example, you can see that the "likely buyers in BigQuery" recipe requires a customers cohort and a connection to BigQuery. Screenshot of a new recipe for likely buyers in BigQuery

  2. In the area for your customers cohort, select it if it already exists in your account and move on to step 4. If not, read the section below.

    Creating a new cohort

    1. If you have a dataset configured, choose an event to define the cohort, then click create cohort and move on to step 4. For customers, you'll select orders, for example. Screenshot of a new cohort in a likely buyers in BigQuery recipe
    2. If you need to create a dataset, select I still need to connect data for this.

      Creating a new dataset

      1. To connect data, choose your method of upload: either a CSV upload, or a connection.
      2. For a CSV upload:
        • Select CSV upload and click next.
        • Choose your CSV from the file picker and click next.
        • In the mapping window, ensure your columns match appropriately in both identity map, where you'll define the people in your data, and event map, where you'll define the events occurring in your data. Screenshot of a new cohort identity map Screenshot of a new cohort event map
        • Click create cohort to confirm the mapping, create the dataset, and create the cohort. This will populate the cohort you created in the customers cohort field.
      3. For a connection:
        • Select a connection and click next.
        • Choose an existing connection, such as your BigQuery instance, and click next.
        • Enter the name of the table where your customer data lives. Screenshot of a new cohort from existing BigQuery connection
        • To create a new connection, choose create a new connection.

        Creating a new connection

        1. To start, read the included setup instructions for your connection type.
        2. Next, fill out your connection parameters. In this BigQuery example, project ID and dataset name are required. Fill these out, then click next. Screenshot of a new cohort from existing BigQuery connection
        3. Enter the name of the table where your customer data lives, and click next.
        4. In the mapping window, ensure your columns match appropriately in both identity map, where you'll define the people in your data, and event map, where you'll define the events occurring in your data.
          Screenshot of a new cohort identity map Screenshot of a new cohort event map
        5. Click create cohort to finish mapping. The cohort will auto-populate in the recipe's cohort field.
  3. In the area for your connection, select your BigQuery connection if you previously configured it, or–if you created a BigQuery connection in the steps to create your customers cohort above, select it.

Screenshot of a filled out recipe for likely buyers in BigQuery

  1. Click proceed to complete the recipe. Your recipe's pipeline will begin processing, and when its status is ready, you'll be able to enable it to sync to your connection, or prepare your CSV for download.