Table of contents

What's new in Watson Studio Local

Check out what's new for Watson Studio Local!

What's new in Watson Studio Local Version 1.2.2 (October 2018)

  • IBM Data Science Experience Local is now IBM Watson Studio.
  • IBM Deployment Manager is now IBM Watson Machine Learning.

Client menu

  • For the Beta remote deployments feature, the dsxr command line utility has been renamed to wsr. Learn more
  • Watson Studio Local now supports Unicode (UTF-8) for data sets.
  • You can now troubleshoot and repair common problems in your Watson Studio Local cluster by running the /wdp/utils/ script. Enter -h for a help page.
  • The Cluster Log page is no longer available in the Admin Console.
  • During the installation process, Watson Studio Local now automatically runs a monitoring tool in the background to collect system performance statistics on every node in the cluster. All data is written to a monitor_results.csv file and graph PNG files located in the installation directory on each node. Learn more
  • Watson Studio Local now supports installing on OpenShift (Beta).

What's new in Data Science Experience Local Version 1.2.1 (August 2018)


  • DSX Local now supports an add-on for Cognos Dashboards Embedded. Learn more

    Cognos dashboard

  • DSX Local now supports an add-on for Watson Explorer oneWEX. Learn more

    Watson Explorer oneWEX

  • DSX Local now supports a Beta add-on for Spark Canvas using either the DSX Local Spark engine or a remote Livy to the Spark engine running on a Hadoop cluster. Learn more

    Spark Canvas


  • From the new Launch Terminal icon ( Launch Terminal), you can open either a Python 2.7 or Python 3.5 Jupyter terminal.

    Jupyter terminal

  • A DSX Local user can now install a library or package in the root conda environment so that it can be shared with other users, and does not get lost when the Jupyter environment restarts. After installing the library or package from a Jupyter notebook or terminal, for example, conda install -y arrow, the user can go to their project Environments panel and click the Save icon (Save) for the running Jupyter environment to save it into a new custom image on the cluster (without having to download or update the entire image). The custom image then appears in the new My Images panel. Any DSX Local user can select this new image by clicking Edit settings next to the corresponding environment. From the admin console, the DSX administrator can see, manage, or validate the images created by DSX Local users in the Image management panel. Learn more
  • A DSX administrator can now push custom images to a Hadoop cluster for remote jobs.
  • When you pull changes and DSX Local detects a Git merge conflict, you can now resolve it in one of two ways:
    Automatic resolution
    Rerun the Git pull but automatically drop either the remote or local changes in a merge conflict.
    Manual resolution
    Resolve the changes manually in either a web terminal or by using your own local merge conflict tool. You can then stage the changed files back to DSX Local.
    Learn more

    Merge conflict

  • From the Git Actions icon ( Git Actions) in the project action bar, Commit and Push are now separate. If you cannot resolve a merge conflict, you can now revert an old commit from the Commit history. Learn more

    Project icons

  • To create library projects, you now add them from the new Libraries page in the DSX Local client. A library owner can now restrict access to it, and grant Viewer privilege to specific DSX Local users. Otherwise, the library defaults to open access. Learn more

    Library viewers

  • You can now click a Test Connection button for a data source to verify that DSX Local can connect to it successfully.
  • While adding a remote data set, you can now browse a schema for most table types. You can also select SQL Query as the object type instead.
  • You can now preview remote data sets for all supported JDBC data sources and custom JDBC data sources by clicking the Preview option.
  • You can now create data sources and remote data sets to access Microsoft SQL Server. Learn more
  • You can now create data sources and remote data sets to access Hyperledger Composer (for building Blockchain applications). Note that the remote data sets can only be added after the data source has been created (rather than during). Learn more
  • Data Refinery automatically opens when you click a data set that supports it (previously you had to select Refine). Additional features:
    • You can now refine data on a remote JDBC or HDFS data set.
    • You can now run Data Refinery jobs on a local data set or a remote HDFS data set.
    • You can now open an information panel by clicking the View info pane button.

    Data Refinery

  • The project page now provides a navigation menu. You can also search, filter, and sort projects.
  • When you search in a project, you can save the web browser URL as a bookmark.
Machine learning
  • You can now create a model group of up to 10 model versions that can be deployed to production at the same time using the same scoring endpoint. Over time, the model versions can be evaluated against each other to assess the best ones to keep on your production environment. Learn more

    Evaluation history by metric

    Evaluation history by date

Model management and deployment

  • DSX Local now requires Deployment Admin permissions to create a project release. The DSX administrator can assign the Deployment Admin permissions to a user. Learn more
  • A Deployment Admin can assign Admin permissions to a member of a project release. The Admin can perform the following tasks:
    • Modify, delete, or launch the project release.
    • Add and administer the members of a project release and assign them permissions.
    • Create, delete, or update asset deployments.
    Learn more
  • A Deployment Admin can assign Developer permissions to a member of a project release. A developer can see the project release and use its existing deployments. For instance, a developer can test the API in a web service and run jobs. A developer can also preview and download the assets of existing deployments.
  • A Deployment Admin can assign Viewer permissions to a member of a project release. A viewer can only see the project release and view its deployments, dashboard, and metrics. A viewer cannot test the API or run jobs.
  • DSX Local now supports Beta remote deployments for web services and script jobs. The Admin of a project release can export the entire release as a TAR.GZ file. Then the release can be uncompressed on a remote node, and its deployments can be started and administered using the new dsxr command line utility. Note that the docker images inside of the project release must be transferred to the remote location manually. Learn more

    Export release

  • In the new Workers tab of a project release, a Deployment Admin can now select and configure workers, including custom images. The Deployment Admin can then select the updated worker when they deploy a job.


  • In the new SSHD service panel in the Admin Console, a DSX administrator can enable SSHD so that users (using a public/public SSH key pair) can ssh directly to the DSX Local cluster through a secure port. This is useful for securely uploading directories or large files to the DSX Local cluster without timing out. Learn more

    SSHD service

  • An administrator on a DSX Local cluster can now define a .condarc file to enforce a common conda policy and configuration across the organization. In the new Conda Management panel in the Admin Console, a DSX administrator can add or remove custom channels that DSX Local users can pull packages from. Learn more

    Conda Management


  • The uninstallation script can now be run as a sudo user. The script also provides a new --keep-docker option to preserve the existing docker installation. Learn more