Use Datasets to Explore data

Dataset is the collection of data models that share the same interests. For example, you are working in an eCommerce company, doing analytics on one standalone model users can only answer certain questions about users (their age, group, gender).

What if you want to know How many orders they have bought, What are the delivery attempts for each user, etc. that when Dataset jumps in, combine all related models (users, orders, order_items, etc.) to serve your data needs. For more detailed information, please refer to our docs: Dataset

Create Dataset

There are 3 approaches to create a Dataset:

  • In Home Screen: You can see a section showing all the recent Datasets you have been using called Explore Datasets. If you don’t have any dataset yet, this section will be empty.

create dataset

  • Left-side bar: You can also see the Create option in folder tree from the left-side bar.

create dataset

  • In a specific Data Model: You can see an option called Explore model. After clicking it, you can see all the datasets using the model you are at

create dataset

After that, a modal jumps out and ask us to fulfill these following information:

  • Dataset name
  • Data Warehouse: The Data Warehouse that stores Data Models that you want to explore
  • Data Models: Select Data models that you want to explore collectively. Please note that they need to have relationships with each other. If you add an isolated model, you cannot use its fields with other models.
  • Relationship: Enable relationships among your selected models. By default, Holistics will automatically detect and disable relationships that introduce ambiguity in the JOIN paths. However, you could also browse the Relationship list and adjust your active relationship as your needs.

After finish setting up, click Save and explore your Dataset.

Here I’m going to create a Dataset:

create dataset

Avoid path ambiguity

As aforementioned, large datasets with multiple models and complex relationship can sometime lead to path ambiguity situation. In this situation, there are multiple possible join paths to combine fields from models, and can produce wrong exploration result.

Holistics will detect such cases and automatically disable some relationships. If the desired path is accidentally disabled, you need to:

  • Find out the possible paths among your models by examining the relationship diagram
  • Deactivate a relationship in the undesired path to cut it off
  • Reactivate the relationships in your desired path

For more information about path ambiguity, please check our docs here.

Explore Dataset

After clicking Save, you will be presented with the Explore screen where you can start dragging in fields, measures, apply conditions, and tweak the visualizations.

explore dataset

When you are satisfied with the exploration result, you can save it as a report inside a Dashboard.

See you in the next guide!