1. Create an account at https://www.mage.ai/sign-up.

    Screen Shot 2022-03-11 at 10.20.55 AM.png

  2. If you already have an account, you can log in at https://www.mage.ai/sign-in.

    Screen Shot 2022-03-11 at 9.56.12 AM.png

  3. Select “Build your first model” from the dashboard.

    Screen Shot 2022-03-11 at 9.29.45 AM.png

  4. Select any of the available use cases (ranking, churn prediction, categorization, and estimation) to start building a machine learning model and enter the model workspace.

    Screen Shot 2022-03-11 at 9.30.39 AM.png

  5. Connect your data to Mage by clicking on “Add data” (⌘+J ****for keyboard shortcut) ****in the model workspace.

    Note: The “Prepare” tab may be renamed as “Train” in the current version of the app.

    Screen Shot 2022-03-11 at 9.32.37 AM.png

    There are several ways to import data into your Mage workspace, but the following methods do not require having your own database or data warehouse set up:

    1. Upload a CSV or JSON file under 100mb. You can find public datasets on Kaggle.

      Screen Shot 2022-03-11 at 10.26.35 AM.png

    2. Clone a public dataset from the sample datasets provided in Mage.

      Screen Shot 2022-03-11 at 9.41.58 AM.png

  6. Prepare your dataset by performing transformer actions on it. Click on the “Edit data” button to see all of the available actions. In the left panel, there may also be suggestions for which actions to use. Applying these transformer actions can help improve the performance of your model.

    Screen Shot 2022-03-11 at 9.36.47 AM.png

  7. Configure your model. Certain inputs are required before training your model. Follow the model input suggestions to begin configuring your model. If an input is optional, you will see a button letting you “Skip suggestion”.

    Screen Shot 2022-03-11 at 10.24.32 AM.png

  8. Train your model. After all suggestions have been completed, you can click on the “Config” or “Start training” buttons to update the model name, modify any model inputs, and begin training your model.

    Screen Shot 2022-03-11 at 10.34.52 AM.png

    If you select the “Standard” model training option, you should receive an email when your model has completed training.

    email.png

  9. Review your model’s performance. After the model is done training, you can review how the model performed, see which features had the greatest influence on your model’s predictions, and look at samples of the predictions.

    Screen Shot 2022-03-11 at 10.36.47 AM.png

    Some useful resources for understanding the model metrics and learning more about the machine learning lifecycle:

  10. Explore the other areas of the app! Mage has other features not mentioned in this guide, so check out the various pages and revisit the app regularly as we are constantly adding new features.

<aside> ❓ Questions or feedback? Contact us.

</aside>