SciServer Compute

SciServer Compute provides Jupyter notebooks in Python or R, and has a SAS Data Volume that can be mounted for access to all SDSS data products.

To create a SciServer account, follow these steps:

  1. Go to www.sciserver.org and click on “Login to SciServer”
  2. Click the “Create a new account” link
  3. Enter a username, a working email address, and a strong password, then agree to the policies and click “Create account”
  4. Check your email and click on the confirmation link
  5. Return to SciServer (you should be automatically logged in), where you will see the SciServer Dashboard

Create a SciServer container for your notebooks:

  1. From the SciServer Dashboard, click on Compute in the second row of icons to open SciServer Compute in a new tab
  2. Click on “Create container” to reserve processing and storage for your work; a new focus window will come up with create container options
  3. Select a Compute Image
    • For a quick start for most uses, choose the SDSS_DR20 or SDSS_DR19 compute image. The images include the sdss_access python package, which is required to access SDSS data, as well as other useful Python libraries such as astropy, astroquery, and specutils.
    • We recommend you select the “Kraken” domain when creating your container. If you plan to use newer tools such as tensorflow (for example if running the machine learning tutorial), you must use this domain.
  4. Give your container a name, and mount the SDSS SAS DR19 or SDSS SAS DR20 Data Volume by clicking its checkbox, then click the Create button. It is important to choose the DR consistently for sdss_access to work correctly. It will be mounted in the Jupyter session under  /home/idies/workspace/sdss_sas_DRXX/.
  5. Click on the name of your newly-created container to open it in a new tab.

Instructions for running DR20 tutorials in SciServer Compute are available here.

Back to Top