Course description:

You will:

  • Learn about DataRobot interface features Data, Models, Insights, Jupyter and Repository
  • How to create a project with your data and how to understand and evaluate the exploratory data analysis so that you can target a feature to predict
  • Check the Leaderboard to see and understand the Models
  • Use insights to understand the features which DataRobot has chosen as those that have the greatest effect.


  • Data Scientists who want to automate the selection of the best predictive model for their Data
  • Business Users who want to perform Data science functions.

No prerequisites are needed.


Tell me, I’ll Forget; Show Me, I’ll Remember; Let Me Do It, I’ll Understand

For each topic:

  • the instructor will explain any new concepts to be used and show how to use DataRobot to achieve the desired result
  • the students will follow along with the instructor
  • the students will then practice what they have learnt by doing the exercises

Course Content:


Here you learn about the industries and some of the types of opportunities that DataRobot can be of assistance, along with the types of predictions that can be made.
Next we look at understanding the User Interface, the type of data that can be used for predicting or forecasting.


Using the Leaderboard you will see the models ranked by the chosen performance metric where you will be able to explore these models.


Allows you to see:

  • Tree-based Variable Importance
  • Hotspots
  • Variable Effects
  • Word Cloud
  • Text Mining

to assist in understanding the model in more detail.


Allows you to use python or r code to run custom models against you data.


Allows you to search through the DataRobot repository for models, where you can preview and run the model as required.

Unlock Holdout

We use DataRobot to unlock the holdout data so that we can check the accuracy of the model’s predictions.


We use DataRobot to train on the initial data and predict on the next set of data.

Time-Series Modelling

We use DataRobot to do time series modelling based on regularly spaced time data.

Alteryx DataRobot Tools

We can both train and predict using Alteryx DataRobot tools.
They interact with DataRobot in the background.

Follow Me

We use the Follow Me sessions to better understand the DataRobot features and answer a series of questions.

  • DataRobot is used to create a new project using patient readmission data, provide a target feature to predict and create the models.
  • Models are used to check the options available to understand the top model and to compare it with a second model.
  • In the Insights session we will explore the insight options to gain more clarity around the models generated by DataRobot.
  • We Unlock Holdout for the patient readmission data and check the accuracy between the validation sample and all of the data including holdout.
  • For the Predict we predict using the training data.
  • The Time-Series option si used against time-series data to see the available options.
  • Using both the Alteryx DataRobot train and predict tools in an Alteryx workflow to interface with DataRobot and to use Tableau to visualise the predicted data.


We use the exercises to reinforce your understand of various DataRobot features and to answer more questions.

  • Data Exercise is used to create a new project, input churn data, check the Feature list, check Frequent Values and check the Table.
  • The Models exercise is used to check the Leaderboard for the top model, check the Learning Curves, check Speed vs Accuracy details and to compare it with a second model.
  • In the Insights exercise we will explore the insight options of Tree-based Variable Importance, Hotspots and Variable Effects.
  • We Unlock Holdout fdata and evaluate the accuracy between the validation sample and all of the data including holdout.
  • We train the model on a year’s data and then predict on the next year’s data.
  • We check Perth weather data for a year using time series to answer questions about the data.


You will:

  • have knowledge of the industries, opportunities and types of predictions that are available in DataRobot will be enhanced.
  • understand the DataRobot interface.
  • use the Leaderboard and the various model features to help your understanding of which model might be best to predict your data.
  • understand the insights that can be found for your model.
  • find out how to use the Jupyter feature of DataRobot.
  • understand how to use the DataRobot model repository.
  • understand how the accuracy of the model can be checked against the holdout sample.
  • be able to both train and predict using DataRobot.
  • understand how time-series works and the options available in DataRobot.
  • be able to use the Alteryx DataRobot tools to both train and predict in DataRobot.


11 Mar 2019


8:00 am - 5:00 pm