5 Step Workflow

Step 1 Exploring Data

Step 1

Exploring Data

Understand the shape, distributions, and quirks of your data before modeling.

Open page →

Step 2 Data Preparation

Step 2

Data Preparation

Clean, transform, and structure the data so it's ready to model.

Open page →

Step 3 Algorithms

Step 3

Algorithms

kmeans tree-based linear-based kNN neural nets SVM naive bayes regularized

Open page →

Step 4 Train

Step 4

Train

Fit the chosen model to the prepared data.

Open page →

Step 5 Evaluation

Step 5

Evaluation

Measure how well the model performs, then iterate.

Open page →

5-step
workflow

The machine learning workflow is comprised of five steps, and often includes iteration after the last step to find optimal solutions.

Click on any of the steps to learn more.

The most important part of this process is something not listed: identifying your workflow’s goal.

This is done by completing relevant background research, talking to stakeholders, and identifying what data you have available to you.

Sometimes goals are not able to come to fruition. This is normal, and requires you to rescope your goal to something that is doable within your time frame.