Datasets

Store, manage, explore and share data in the Polyteia Platform.

Datasets are a central resource in the Polyteia Platform and serve as the fundamental module for storing data in tabular form. This data can be explored and used in other resources. Users can import external data through manual file uploads, access keys (PAK), or forms.

Dataset administrators manage data access by assigning roles, such as editors. With appropriate access, users can generate insights and reports from one or more datasets.

Features

• Upload and store data Upload CSV, Excel, JSON, or Parquet files to store structured data.

• Schema editing Rename columns and change data types under the Details tab.

• Tags and documentation Add context and categorization directly on the overview page.

• Explore and visualize Open a dataset in the editor to analyze it or create insights.

Accessing datasets

From a solution

1. Go to Home → Workspace → Solution.

2. Click on the Datasets tab.

3. Select a dataset to view or edit.

How to create a new dataset from the solution:

• From the solution overview: Click on Create dataset

• From the Datasets tab: Click on + New dataset

From the home page

1. Select datasets in the left page navigation on the home page.

2. Click on a dataset title.

---

Managing datasets

Click on the ... menu at the top right of the dataset overview page to access these actions with the appropriate permissions:

• Explore data - opens the dataset in table view

• Edit dataset details - change title, source, or description

• Upload data - replace or add to existing data

• Share - manage sharing options

• Duplicate dataset - create a copy of the dataset

• Download as - export the dataset in a supported format

• Copy dataset ID - copy the unique identifier to clipboard

• Delete dataset - permanently delete the dataset (owners only)

Roles and permissions

Datasets in the Polyteia Platform have three roles:

• Owners have full control, including managing settings, sharing access, and sharing or transferring ownership.

• Editors can modify data but cannot change dataset settings or permissions.

• Viewers can only see the dataset's metadata, not its content.

All solution members can create datasets. The creator of a dataset is the default owner but can share access or transfer ownership if needed.

Insight permissions are inherited from dataset roles, meaning only dataset owners and editors can create and run insights, while viewers can only see dataset and insight metadata.

Dataset permissions

Permissions in datasets determine who can create, modify, share, and analyze data. All solution members can create datasets, and the creator automatically becomes the default owner with the ability to share or transfer ownership if needed. These permissions also directly affect insights, as only dataset owners and editors can create and run insights, while viewers can only see dataset and insight metadata.

Permission	Dataset Owner	Dataset Editor	Dataset Viewer
Edit dataset metadata (e.g., name, schema)	✓	✕	✕
Share dataset and transfer ownership	✓	✕	✕
Delete dataset	✓	✕	✕
Duplicate dataset	✓	✓	✕
View, add, modify, and delete data	✓	✓	✕
Run queries & generate insights	✓	✓	✕
View dataset metadata	✓	✓	✓

How dataset permissions control insights

Insights inherit permissions from the datasets they are based on, meaning dataset owners and editors can create, edit, and run insights, while dataset viewers have limited or no access. Since dataset permissions determine who can interact with insights, users must have the appropriate dataset role to generate or modify insights.

However, administrative actions for insights — such as renaming or deleting — are reserved for the insight owner (the creator of the insight). Insight ownership cannot be transferred, but users can duplicate an insight to create a new version under their own control.

Best practices

• Assign editor roles carefully to prevent unwanted data changes.

• Use viewer roles for users who only need to know the metadata and schema of the dataset.

• If another user needs full control, make the user an additional owner of the dataset.