About assets

An asset is an item of data or data analysis in a project or catalog. Data assets contain data, for example, a file or a data set that’s accessed through a connection to an external data source. Analytical assets run code to analyze data.

An asset has metadata that you see in the project or catalog. An asset might also have associated files or assets, or require specific IBM Cloud services. For example, many types of analytical assets require a service.

The information that you can view about an asset and the actions that you can perform on an asset vary by asset type and whether the asset is in a project or in a catalog. A few actions apply to all asset types, for example, you can create, edit the properties of, and remove or delete all types of assets. You can edit the contents of most types of assets in projects. For data assets that contain relational data, you can shape and cleanse the data with the Data Refinery tool. For analytical assets, you can edit the contents in the appropriate editor tool. In catalogs, you can edit only the metadata for the asset.

Data assets:

Analytical assets:

Not all asset types are supported in catalogs.

Asset type In projects? In catalogs?
Data asset from a file
Connected data asset
Folder asset
Connection asset
Notebook asset
Dashboard asset
Watson Machine Learning model asset
Synthesized neural network asset  
Experiment asset  
Visual Recognition model asset  
Natural Language Classifier model asset  
Modeler flow asset  
Data Refinery flow asset  
Streams flow asset  


Data asset from a file

A data asset from a file is a pointer to an uploaded file. You create a data asset from a file by uploading a file from your local system into a project or a catalog. The file is stored in the project’s or catalog’s object storage. The contents of the file can include structured data, unstructured textual data, images, and other types of data. You can create a data asset with a file of any format, however, you can do more actions on CSV files than other file types.

In projects, you can see data assets from a file in the Data Assets section on the Assets page.

Information you can see about data assets from files

In projects, you can see this information about data assets from files:

  • The asset name and description
  • The tags for the asset
  • The name of the person who created the asset
  • The size of the data
  • The date when the asset was added to the project
  • The date when the asset was last modified
  • A preview of the data, for CSV, Avro, Parquet, Microsoft Excel, PDF, text, and image files
  • A profile of the data, for CSV, Avro, Parquet, Microsoft Word, PDF, text, and HTML files
  • The lineage

In catalogs, you can see this information about data assets from files:

  • The asset name and description
  • The tags for the asset
  • The date when the asset was added to the catalog
  • The connection to IBM Cloud Object Storage for the associated file for the asset
  • The classification of the asset, if policies are enabled for the catalog
  • A preview of the data, for CSV, Avro, Parquet, Microsoft Excel, PDF, text, and image files
  • A profile of the data, for CSV, Avro, Parquet, Microsoft Word, PDF, text, and HTML files
  • The lineage
  • The owner, the privacy setting, and the members of the asset
  • Ratings and reviews

Associated files or assets for data assets from files

A data asset that you create from a file has an associated file in the object storage for the project or catalog.

Associated services for data assets from files

A data asset from a file requires the IBM Cloud Object Storage service instance that you specify when you create a project or catalog.

Actions on data assets from files

You can perform these actions on data assets from files in projects:

  • Create in a project
  • Analyze CSV, Avro, or Parquet files in analytic assets
  • Refine Avro, CSV, JSON, Parquet, or text files with the Data Refinery tool
  • Annotate files by using a third-party crowd annotation platform
  • Edit asset properties
  • Download the file that is associated with the asset to your local system
  • Publish a copy of the asset from a project to a catalog and add a copy of the associated file to the catalog object storage
  • Anonymize CSV, Avro, or Parquet files to mask or substitute sensitive data in a column with a policy
  • Remove from a project

You can perform these actions on data assets from files in catalogs:

  • Create in a catalog
  • Edit asset properties
  • Download: download the file that is associated with the asset to your local system
  • Copy to projects: add a copy of the asset from the catalog to the project and add a copy of the associated file to the project object storage
  • Anonymize (CSV, Avro, or Parquet files): mask or substitute sensitive data in a column with a policy
  • Rate and review an asset in a catalog
  • Control access to the asset
  • Classify: specify a classification for the asset, if policies are enforced in a catalog
  • Remove from a catalog

Connected data asset

A connected data asset is a pointer to data that is accessed through a connection to an external data source. You create a connected data asset by specifying a relational table or view, a set of partitioned data files, or a file, and the connection asset. When you access a connected data asset, the data is dynamically retrieved from the data source.

In projects, you can see connected data assets in the Data Assets section on the Assets page.

Data connections marked with a key icon are locked. If you are authorized to access the data source, you are asked to enter your personal credentials the first time you select it. This is a one-time step that permanently unlocks the connection for you. After you have unlocked the connection, the key icon is no longer displayed. See Adding connections to projects. Connections with personal credentials are already unlocked if you created the connections yourself.

Partitioned data

You can create a connected data asset from partitioned data files. Partitioned data assets have previews and profiles and can be anonymized like relational tables. However, you cannot yet shape and cleanse partitioned data assets with the Data Refinery tool.

Partitioned data is recognized and treated like a relational table if the files meet these requirements:

  • The files have a prefix of part-.
  • The files are in a single folder within IBM Cloud Object Storage that contains no other files.

Information you can see about connected data assets

In projects, you can see this information about connected assets:

  • The asset name and description
  • The tags for the asset
  • The name of the person who created the asset
  • The size of the data
  • The date when the asset was added to the project
  • The date when the asset was last modified
  • A preview of relational data
  • A profile of relational data
  • The lineage

In catalogs, you can see this information about connected data assets:

Associated files or assets for connected data assets

A connected data asset requires an associated connection asset, which provides the information on how to connect to the appropriate data source.

Associated services for connected data assets

A connected data asset does not use additional services.

Actions on connected data assets

You can perform these actions on connected data assets in projects:

  • Create in a project
  • Analyze (relational tables or CSV, Avro, or Parquet files): analyze in analytic assets
  • Refine relational tables or Avro, CSV, JSON, Parquet, or text files with the Data Refinery tool
  • Edit asset properties
  • Publish a copy of the asset from a project to a catalog and add a copy of its associated connection asset to the catalog
  • Anonymize relational tables or CSV, Avro, or Parquet files to mask or substitute sensitive data in a column with a policy
  • Download the asset as a file to your local system
  • Delete from a project

You can perform these actions on connected data assets in catalogs:

  • Create in a catalog
  • Edit asset properties
  • Copy to projects: add a copy of the asset from the catalog to the project and add a copy of its associated connection asset to the project
  • Anonymize relational tables or CSV, Avro, or Parquet files to mask or substitute sensitive data in a column with a policy
  • Rate and review
  • Control access to the asset
  • Classify the asset, if policies are enforced in a catalog
  • Download the asset as a file to your local system
  • Delete from a catalog

Connection asset

A connection asset is the information necessary to create a connection to a data source. You create a connection asset by providing the connection information.

In projects, you can see connection assets in the Data Assets section on the Assets page.

Information you can see about connection assets

You can see this information about connection assets in projects:

  • The name of the connection
  • The description of the connection
  • Connection details, such as the host name, port number, user name, and password

You can see this information about connection assets in catalogs:

  • The name of the connection
  • The description of the connection
  • The tags for the asset
  • The date when the asset was added to the catalog
  • The data source type of the connection
  • The classification of the asset, if policies are enabled for the catalog
  • Connection details, such as the host name, port number, user name, and password
  • The owner, the privacy setting, and the members of the asset
  • Ratings and reviews

Associated files or assets for connection assets

Connection assets are necessary for connected data assets, folder assets, and Data Refinery flow assets.

Associated services for connection assets

A connection asset can use a data service as a data source. A connection asset for an on-premises data source requires a Secure Gateway service instance.

Actions on connection assets

You can perform these actions on connection assets in a project:

  • Create in a project
  • Discover: automatically add all tables and views from a connection asset to a project
  • Publish to a catalog: add a copy of the asset from the project to a catalog
  • Edit asset properties
  • Edit connection values
  • Delete from a project

You can perform these actions on connection assets in catalogs:

  • Create in a catalog
  • Discover: automatically add all tables and views from a connection asset to a project
  • Edit asset properties
  • Edit connection values
  • Copy the asset from the catalog to the project
  • Rate and review
  • Control access to the asset
  • Classify the asset, if policies are enforced in a catalog
  • Delete from a catalog

Folder asset

A folder data asset is a pointer to a folder in IBM Cloud Object Storage. You create a folder asset by specifying the path to the folder and the IBM Cloud Object Storage connection asset. You can view the files and subfolders that share the path with the folder asset. The files that you can view within the folder asset are not themselves data assets. For example, you can create a folder asset for a path that contains news feeds that are continuously updated.

In projects, you can see folder assets in the Data Assets section on the Assets page.

Information you can see about folder assets

You can see this information about folder assets in projects:

  • The name of the person who created the asset
  • The size of the folder
  • The date when the asset was added to the project
  • The date when the asset was last modified
  • A preview of the contents of the folder

You can see this information about folder assets in catalogs:

Associated files or assets for folder assets

A folder asset requires an associated connection asset, which provides the information on how to connect to the appropriate data source.

Associated services for folder assets

A folder asset does not use additional services.

Actions on folder assets

You can perform these actions on folder assets in projects:

  • Create in a project
  • Create connected data assets from the files within the folder asset
  • Edit asset properties
  • Refine Avro, CSV, JSON, Parquet, or text files within the folder asset with the Data Refinery tool
  • Publish a copy of the asset from a project to a catalog and add a copy of its associated connection asset to the catalog
  • Download the contents of the folder asset as a set of files to your local system
  • Remove from a project

You can perform these actions on folder assets in catalogs:

  • Create in a catalog
  • Create connected data assets from the files within the folder asset
  • Edit asset properties
  • Copy the asset from the catalog to the project and add a copy of its associated connection asset to the project
  • Download the contents of the folder asset as a set of files to your local system
  • Rate and review
  • Control access to the asset
  • Classify the asset, if policies are enforced in a catalog
  • Remove from a catalog

Notebook asset

A notebook asset is based on a Jupyter notebook file. You create a notebook asset in the Jupyter notebook editor tool by coding or importing a notebook file. In a notebook, you can run small pieces of code that process your data and immediately view the results.

Information you can see about notebook assets

You can see this information about notebook assets in project:

  • The name of the asset
  • The description of the asset
  • The name of the person who last edited the asset
  • The date when the notebook asset was created in the project
  • The date when the asset was last modified
  • The programming language of the notebook
  • Whether a read-only version of the notebook is shared
  • The runtime status of the notebook
  • Whether the notebook is scheduled to run
  • The contents of the notebook
  • The comments added to the notebook

You can see this information about notebook assets in catalogs:

Associated files or assets for notebook assets

A notebook asset has an associated .ipynb file in the object storage for the project or catalog.

You can use data assets from a file or connected data assets within a notebook. You can also access data in other ways.

Associated services for notebook assets

A notebook asset requires the IBM Cloud Object Storage service instance that you specify when you create a project.

You can choose to associate a notebook with a Spark service instance. However, you can associate a notebook with a runtime environment, which does not require a service.

If you want to train a model in a notebook using Watson APIs, use the appropriate Watson service.

Actions on notebook assets

You can perform these actions on notebook assets in projects:

  • Create in a project
  • Publish a copy of the asset from a project to a catalog
  • Publish a copy of the notebook file to GitHub or as Gist
  • Share a read-only version of the notebook outside of Watson Studio
  • Download the notebook as an .ipynb file
  • Schedule: specify a time or recurring times to run the notebook
  • Duplicate the notebook in a project
  • Edit asset properties
  • Edit the notebook contents in the Jupyter notebook editor
  • Run the notebook code cells
  • Create multiple versions of the notebook
  • Comment on the notebook
  • Remove from the project

You can perform these actions on notebook assets in catalogs:

  • Copy the asset from the catalog to the project
  • Download the notebook as an .ipynb file
  • Edit asset properties
  • Rate and review
  • Control access to the asset
  • Classify the asset, if policies are enforced in a catalog

Dashboard asset

A dashboard asset is a set of visualizations of analytical results. You create a dashboard asset in the dashboard editor tool without writing code.

Information you can see about dashboard assets

You can see this information about dashboard assets in projects:

  • The name of the asset
  • The description of the asset
  • The name of the person who created the asset
  • The date when the asset was created
  • The name of the person who last edited the asset
  • The date when the asset was last edited
  • The name of the associated Cognos Dashboard Embedded service
  • The dashboard contents

You can see this information about dashboard assets in catalogs:

Associated files or assets for dashboard assets

A dashboard asset has an associated JSON file in the object storage for the project.

You must use data assets from a file or connected data assets with a dashboard to provide the data to display.

Associated services for dashboard assets

A dashboard requires a Cognos Dashboard Embedded service instance.

Actions on dashboard assets

You can perform these actions on dashboard assets in projects:

  • Create in a project
  • Share a read-only version of the dashboard outside of Watson Studio
  • Publish a copy of the asset from a project to a catalog
  • Download the dashboard as a JSON file
  • Duplicate the dashboard in a project
  • Edit asset properties
  • Edit the dashboard definition in the dashboard editor tool
  • Remove from the project

You can perform these actions on dashboard assets in catalogs:

  • Copy the asset from the catalog to the project
  • Edit asset properties
  • Download the dashboard as a JSON file
  • Rate and review
  • Control access to the asset
  • Classify the asset, if policies are enforced in a catalog

Watson Machine Learning model asset

A Watson Machine Learning model asset is based on machine learning algorithms that are optimized for a training data set. You create a Watson Machine Learning model asset when you create a model with the Model Builder, Flow Editor, or Experiment Builder tool. You train a model by running it against a data set until its predictions are accurate.

Information you can see about Watson Machine Learning model assets

You can see this information about Watson Machine Learning model assets in projects:

  • The name of the asset
  • The description of the asset
  • The date when the model was last trained
  • The name of the associated Watson Machine Learning service
  • The name of the associated Spark service
  • The status of the model
  • The model type
  • The label column
  • The latest version
  • The input schema
  • The last evalutation result
  • The versions of the model
  • The model deployment name, status, and type
  • The lineage

You can see this information about trained Watson Machine Learning model assets in catalogs:

Associated files or assets for Watson Machine Learning model assets

You use data assets from a file or connected data assets to train a Watson Machine Learning model.

If you configure performance monitoring, you need a connection to an IBM Db2 Warehouse on Cloud data source.

Associated services for Watson Machine Learning model assets

Watson Machine Learning model assets require a Watson Machine Learning service instance.

The Model Builder tool requires a Spark service instance. You can use other methods to create a model asset.

If you configure performance monitoring, you need an IBM Db2 Warehouse on Cloud service instance.

Actions on Watson Machine Learning model assets

You can perform these actions on trained Watson Machine Learning model assets:

  • Create in a project
  • Edit asset properties
  • Test the model
  • Deploy: save the model to the model repository that is associated with your Watson Machine Learning service
  • Edit the model in the appropriate model tool
  • Create multiple versions of the model
  • Monitor the performance and automatically retrain and redeploy the model
  • Download the model as a PMML (.xml) file
  • Publish a copy of the asset from a project to a catalog
  • Remove the model from a project
  • View the deployment of the model
  • Delete the deployment of the model

You can perform these actions on trained Watson Machine Learning model assets in catalogs:

  • Edit asset properties
  • Copy the asset from a catalog to a project
  • Rate and review
  • Control access to the asset
  • Classify the asset, if policies are enforced in a catalog
  • Remove the model from a catalog

Synthesized neural network asset

The NeuNetS tool in IBM Watson Studio synthesizes a neural network and trains it on your training data without you having to design or build anything by hand. In the NeuNetS tool, you can view or download performance metrics, including statistics about classes and a confusion matrix showing how well the model is performing.

Information about the asset

You can see the following information about synthesized neural network assets in projects:

  • Name of the synthesized neural network
  • Content type for the asset
  • Date when the asset was last edited

Associated files or assets for synthesized neural network assets

You must use data assets from a file or connected data assets to create a synthesized neural network. Synthesized neural network assets also require an IBM Cloud Object Storage service instance to store training results.

Associated services for synthesized neural network assets

A synthesized neural network asset requires these services:

  • IBM Watson OpenScale
  • IBM Cloud Object Storage
  • IBM Watson Machine Learning

Actions on synthesized neural network assets

You can perform these actions on synthesized neural network assets in projects:

  • Create in a project
  • Edit the asset in the NeuNetS tool
  • Vew performance metrics and statistics
  • Save as a Watson Machine Learning model asset
  • Synthesize a neural network
  • Deploy or download the trained model
  • Delete the asset from a project

Experiment asset

An experiment asset is based on a logical grouping of one or more model training definitions. You create an experiment asset in the Experiment Builder tool. A deep learning experiment runs a set of models to identify the best combination of data in conjunction with hyperparameters to optimize the performance of your neural networks. You can also create and run deep learning experiments with other methods that do not result in experiment assets in Watson Studio.

Information you can see about experiment assets

You can see this information about experiment assets in projects:

  • The name of the asset
  • The description of the asset
  • The name of the person who created the asset
  • The date when the asset was last edited
  • The training definition
  • The experiment framework and execution command
  • The experiment training logs
  • Comparisons of experiment runs
  • Metrics about experiment runs

Associated files or assets for experiment assets

An experiment asset requires a training definition file and generates training log files in the project’s object storage.

Associated services for experiment assets

Experiment assets require a Watson Machine Learning service instance. Experiment assets also require a IBM Cloud Object Storage service instance to store training definition files and training log files.

Actions on experiment assets

You can perform these actions on experiment assets:

  • Create in a project
  • Edit asset properties
  • Edit the experiment in the Experiment Builder tool
  • Run an experiment
  • Save as a Watson Machine Learning model asset
  • Remove from a project

Visual Recognition model asset

A Visual Recognition model asset is based on a set of custom image classifiers that you define. You create a Visual Recognition model asset with the Visual Recognition model builder tool. You can also create Visual Recognition models outside of Watson Studio.

Information you can see about Visual Recognition model assets

You can see this information about Visual Recognition model assets in projects:

  • The name of the asset
  • The description of the asset
  • The name of the person who created the asset
  • The date when the asset was last edited
  • The training data
  • The classifications

Associated files or assets for Visual Recognition model assets

Visual Recognition model assets require archive files containing image files that are uploaded as data assets.

Associated services for Visual Recognition model assets

Visual Recognition model assets require a Visual Recognition service instance and the IBM Cloud Object Storage service that you specify when you create a project to store image files.

Actions on Visual Recognition model assets

You can perform these actions on Visual Recognition model assets:

  • Create in a project
  • Edit asset properties
  • Test in the Visual Recognition model builder tool
  • Edit the model in the Visual Recognition model builder tool
  • Retrain the model
  • Delete from a project

Natural Language Classifier model asset

A Natural Language Classifier model asset is based on a set of custom textual content classifiers that you define. You create a Natural Language Classifier model asset with the Natural Language Classifier model builder tool.

Information you can see about Natural Language Classifier model assets

You can see this information about Natural Language Classifier model assets in projects:

  • The name of the asset
  • The description of the asset
  • The name of the person who created the asset
  • The date when the asset was last edited
  • The training data
  • The classifications

Associated files or assets for Natural Language Classifier model assets

Natural Language Classifier model assets use CSV files containing text strings that are stored as data assets. You can upload CSV files into the project’s object storage, or paste text strings into the Natural Language Classifier model editor tool, which automatically saves text strings as data assets.

Associated services for Natural Language Classifier model assets

Natural Language Classifier model assets require a Natural Language Classifier service instance and the IBM Cloud Object Storage service that you specify when you create a project to store CSV files that contain text strings.

Actions on Natural Language Classifier model assets

You can perform these actions on Natural Language Classifier model assets:

  • Create in a project
  • Edit asset properties
  • Test in the model builder tool
  • Edit the model in the model builder tool
  • Retrain the model
  • Delete from a project

Modeler flow asset

A modeler flow asset is based on a graphical representation of a data model or a neural network design. You create a modeler flow asset with the Flow modeler tool. You can create a machine learning flow or a deep learning flow. A machine learning flow is a graphical representation of a data model. A deep learning flow is a graphical representation of a neural network design, which you can use to design and run experiments.

Information you can see about modeler flow assets

You can see this information about modeler flow assets in projects:

  • The name of the asset
  • The description of the asset
  • The name of the person who created the asset
  • The date when the asset was last edited
  • The type of modeler flow
  • Comments on the modeler flow
  • Annotations in the modeler flow
  • Versions of the modeler flow
  • The nodes and operations of the modeler flow
  • The SPSS visualizations, for SPSS nodes
  • The output of modeler flow runs
  • The versions of the modeler flow

Associated files or assets for modeler flow assets

You must use data assets from a file or connected data assets to create a modeler flow.

Associated services for modeler flow assets

Modeler flow assets require a Watson Machine Learning service instance.

Actions on modeler flow assets

You can perform these actions on modeler flow assets:

  • Create in a project
  • Edit asset properties
  • Edit the modeler flow in the Flow editor tool
  • Run the modeler flow
  • Save a version of the modeler flow
  • Comment on or annotate a modeler flow
  • Export the flow as an SPSS Modeler Stream file (.str)
  • Publish as a training definition for deep learning experiments
  • Delete from a project

Data Refinery flow asset

A Data Refinery flow asset is based on an ordered set of steps to cleanse, shape, and enhance data. You create a Data Refinery flow asset with the Data Refinery tool.

Information about the asset

You can see this information about Data Refinery flow assets in projects:

  • The name of the asset
  • The description of the asset
  • The name of the person who created the asset
  • The date when the asset was last edited
  • The steps in the Data Refinery flow
  • A preview of the source and target data sets
  • Information about each Data Refinery flow run
  • The schedule for running the Data Refinery flow

Associated files or assets for Data Refinery flow assets

You must use data assets from a file or connected data assets to create a Data Refinery flow.

Associated services for Data Refinery flow assets

A Data Refinery flow asset does not use additional services directly.

Actions on Data Refinery flow assets

You can perform these actions on Data Refinery flow assets in projects:

  • Create in a project
  • Edit asset properties
  • Edit the Data Refinery flow in the Data Refinery tool
  • Run the Data Refinery flow
  • Visualize the data
  • Copy data from a source to a target
  • Schedule: specify a time or recurring times to run the Data Refinery flow
  • Delete from a project

Streams flow asset

A streams flow is a continuous pipeline of streaming data that includes operators to ingest, transform or score, and store. You create a streams flow by using an out-of-the-box example flow, the wizard, or directly in the canvas.

Information about the asset

You can see the following information about stream flows in projects:

  • Name of the streams flow
  • Name of the person who modified the asset
  • Date when the asset was last edited

Associated files or assets for streams flow assets

A streams flow typically references connection assets, which provides information on how to access data sources and targets.

Associated services for streams flow assets

A streams flow asset requires an instance of the Streaming Analytics service.

Actions on streams flow assets

You can perform these actions on streams flow assets in projects: