First Steps

Once you’ve registered your account, there are a few steps to take before we start training. Here is a breakdown of the steps you’ll need to take.

Step 1: Generate a Schema

The schema works as a guide for your machine learning model by describing which data points it should learn from. It points out which data points are important to watch out for (objects) and how they can be used by your customers (interactions).
To generate a schema, click Create A Data Schema from the Setup page:
Screenshot highlighting a button labelled “Create A Data Schema”
Screenshot highlighting a button labelled “Create A Data Schema”
This takes you to Begin’s schema IDE. From here, paste your schema in the textbox below or click Upload Schema json File:
notion image
The Schema editor will check for any errors in the file and flag any changes that need to be made.
For more information on how to create a schema, read our article on schema generation guidelines.

Step 2: Process the Data

Once the schema has been saved, you’ll be taken to your Integration Code.
notion image
  1. Select the tab for the software code you’d like to integrate with. We currently work with Android, iOS, and Python-based software.
  1. Click Copy Code.
  1. Paste the SDK into your codebase.
  1. Click Finish
Once your software runs, Begin will process locally, on your device, the data points you’ve outlined in the schema, including:
  • User information
  • Other objects
  • User interactions with the objects
WHEN DO I MAKE THESE CALLS? Generally, you should trigger these commands every time there is a change in user attributes, object attributes, or new interactions.

Step 3: Create A Project With An Objective

Once you’ve pasted your integration code and clicked Finish, the next step is to create a new project.
A Project takes the data points listed in your schema and maps them to an advanced machine learning algorithm that learns from the users patterns for your specific objective.
To make a project:
  1. Enter a name for your project.
  1. Include a brief description (optional).
  1. From the dropdown, select an objective (this determines the learning algorithm to use).
  1. Choose a schema. Begin will scan the objects within the schema and populate a dropdown menu.
  1. Select the object from the “object to be used” dropdown.
notion image
FOR EXAMPLE: I’m building an app called Book Club, that matches books to interested readers. I’m looking to use Begin’s platform to help groups find their next read. 1. I name the project “Book Club Recommendations”, then provide a description. 2. I choose “recommendation” as my project objective. 3. I select the “book club” schema I uploaded earlier during setup. 4. Finally, I select “book” as the object to be used.

Step 4: Call the Project

From your IDE, initiate a call to your project using the commands found in the SDK index linked below:
SDKs Index
Once entered, Begin will push a mathematical hash representing user fields, object fields, and interactions to the algorithm’s server. From here, it will learn and provide relevant insights based on your project’s objective.
We recommend calling the project regularly, to account for changes in your user base and how they interact with your software.