Skip to content
Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?
Go to file
Cannot retrieve contributors at this time

Publishing on the Regional Data Center's Open Data Portal

This section of the publisher's guide covers the process for manually publishing a dataset to the Regional Data Center's open data portal.

The open data portal used by the Regional Data Center is powered by open source software called CKAN.

Before creating a dataset the first time, you'll need to make sure that:

  1. You are logged in to CKAN and your account has the appropriate permissions enabling you to publish.
  2. Your data file is machine-readable and is in an appropriate file format. Refer to the "preparing your data" section of this guide.
  3. You have created a data dictionary for your file. See the Creating a data dictionary section of this guide.
  4. You are prepared to create a metadata record for your dataset. Metadata is created as part of this publishing process. Details in how to do this are included below.

Before updating an existing resource within the dataset, you'll need to make sure that:

  1. You are logged in to CKAN and your account has the appropriate permissions enabling you to publish data.
  2. Your data file is machine -eadable and is in an appropriate file format. Refer to the "preparing your data" section of this guide.

Creating a CKAN Account

Creating an account on CKAN involves visiting the Regional Data Center web site and entering a user name, password, and email address.

Regional Data Center staff or your organization's administrator will also associate your CKAN acount with the appropriate organization.

Logging in to CKAN

To be able to publish a dataset, you must be logged in to the open data portal using your CKAN account. Your account must also be provided with editor or admin priveliges, as described below. To login, please visit the Regional Data Center's login page.

Accounts and Roles

Accounts in CKAN can have one of three roles. The roles available within CKAN include:

  • member: members can view private datasets owned by their organization, but cannot add data to the site.
  • editor: can edit, read and create new objects on CKAN
  • admin: admin can do anything including: edit, read, delete, and update permissions (change authorizations for that object)

Assigning Permissions

All accounts initially created on CKAN lack publishing permissions. To publish data, your account will need to be assigned either an editor or admin role.

Only the Regional Data Center or the designated CKAN administrator at your organization is able to assign the appropriate role to your account.

You will need to provide them with your CKAN user name to assign you as an administrator or editor. Please contact the Regional Data Center if you have any questions.

A CKAN administrator in your organization (or WPRDC staff) will also need to provide your account with permission to add data to topical categories or groups on the web site. The administrator will use the group adder tool to assign these permissions.

Within an organization, we'd like there to be a primary point of contact that can coordinate among all publishers within your organization, if applicable.

Datasets and Resources Defined

  • It's best to think of datasets within CKAN as a container that holds information about the dataset itself in a metadata record, along with a number of resources.

  • Resources in CKAN are items within the dataset "container" itself. Resources can include files of many data types within the same dataset, including data file formats, PDFs, images, hyperlinks, and text documents. Resources in the same dataset might contain the data for the same information broken into a separate resource by year, a data dictionary describing the fields in a data table, or a hyperlink to a dataset or tool hosted in a different location. CKAN accepts data in many different formats. A resource can be in a variety of formats, including (but not limited to): CSV, Excel spreadsheet, text file, JSON, PDF document, JPEG image file, etc. CKAN can store the resource internally or provide a link to resources hosted elsewhere on the web.

Creating a Dataset

Manually uploading data to CKAN involves a two-step process. The first step in the process involves the creation of a dataset. The second involves the uploading of resources.

To start the process, assuming you're logged in to CKAN, click on the "Add Dataset" button from the Dataset web page: alt text


The first step in the process of creating a dataset involves the creation of a metadata record. Metadata is a structured framework for documenting data. The metadata standard we’re using is largely based on one used by San Francisco and U.S. Were also grateful for advice from Digital Scholarship Services at the University of Pittsburgh

We’ve programmed the open data portal software to automatically complete some of the metadata. We’ve also provided drop-down forms saving you from having to type pre-defined responses where possible.

We also don’t require all of the metadata elements to be completed when data is loaded to the portal. We have a few required fields (as noted below), and more detail can be added to the metadata later.

You may find it convenient to use our metadata entry worksheet prior to creating a dataset and metadata record on the web site. It will take some planning to gather all necessary materials for the metadata record. If the metadata information is gathered in advance, It may take as littla as five minutes to enter a metadata entry for a dataset. It shouldn’t need to be updated unless some of the details have changed.

Title (Required)

Short human-readable name of the asset. Should be in plain English and include sufficient detail to facilitate search and discovery. Avoid acronyms. Use "title case" spelling, and no need to list dates. Don't include the organization name here.


URL link to the dataset landing page on the open data portal. This field is automatically populated by the software.

Description (Required)

Provide a longer description of the data (compared to the title) written in plain language that can be readily understood by non-technical users. It's a good practice to include your organization's name in the description

Tags (Required)

Keywords that describe the dataset. Enter separated by semicolons. Acronyms acceptable. Use technical and non-technical terms. Use as many as needed. Use plural forms of the word, and also singular forms if different from plural (i.e. leaf/leaves, wife/wives).

License (Required)

License definitions and additional information on licenses can be found on the Creative Commons web site and the Open Definition web site. A Public Domain Dedication in lieu of a license is also a choice that will allow for maximum re-use of your data if no copyright is present. This information is entered using a drop-down menu.

Group/Topic (Required)

The group/topic of the dataset identified by the list of possible values. If a data set can fall into multiple categories, select the one which is most-appropriate. This information is entered using a drop-down menu.

Organization (Required)

Name of the organization sharing data. This information is entered using a drop-down menu.


Name of the data source department or division (if applicable). This information is entered using a drop-down menu. If no choices are provided, please leave it blank.

Access Level

Public datasets are visible to all, and private datasets are only visible to other users in the same organization. New datasets should be set as Private at the time of publication. The Regional Data Center will review the data as a final privacy check, and will set it as Public if no privacy issues are present. This information is entered using a drop-down menu.

Public Access Level Comment

An explanation of any general steps that have been taken to make a sensitive dataset public, where appropriate, including obfuscation, aggregation, or anonymization.

Temporal Coverage

Start/End periods covered by the data. Separate start and end by a "/". Use years (YYYY), dates (YYYY-MM-DD), or dates and times (YYY-MM-DD"T"HHMMSS). If the data is a snapshot from a particular date or year, please omit the "/". Examples: 4 digit year range (2013/2015), date range (2011-02-14/2013-07-04), date/time range (2011-02-14T12:00:00Z/2013-07-04T19:34:00Z), 1 year (2015), single date (2/16/2016).

Geographic Unit

At what geographic unit is the data collected? For example, if the data is collected by address, it would be Street Address. This information is entered using a drop-down menu.

Data Notes

Are there any concerns about overall data reliability? Are there any changes in data collection or methods that the user should be aware of? Are there any constraints with data accuracy? What levels of confidence with this dataset could the user reasonably assume?

Related Documents

Related documents such as technical information about a dataset, developer documentation, URL, etc.

Frequency - Data Change (Required)

Frequency with which dataset changes. This can be a relative frequency - If the file changes every 1st and 4th Monday, we can code this as bi-monthly, etc. This information is entered using a drop-down menu.

Frequency - Publishing (Required)

Frequency with which dataset is published. This information is entered using a drop-down menu.

Data Steward Name (Required)

Data steward's name. Who manages the data and is responsible for making changes to the data? Who understands what the dataset includes and can answer questions about it?

Data Steward Email (Required)

Data steward's email address.

We recorded a video to show how the process works...

alt text

Data Dictionary

We also strongly encourage publishers to provide a data dictionary for each data table. Data dictionaries help data users understand the underlying structure of the data file. The data dictionary can be created in a spreadsheet, saved in .csv format, and uploaded to the dataset as a resource. For an example, please see the example included with the Allegheny County Property Assessment data.

Our system now supports integrated data ditionaries, making it unneccessary to upload a separate CSV file. More on integrated data dictionaries here.

This data dictionary should include (at minimum):

  • Name of each field
  • Description of the data field

The data dictionary can also include information about each field in the data:

  • Data types (text, numeric, boolean, etc.)
  • Format (special details related to the format, including currency, decimal placement, dates/times
  • Field length (# of characters)
  • Unique ID/Primary Key
  • Permitted/prohibited values
  • Public access limitations

You can read more about data dictionaries, how to make them, and our suggested formats here.

Creating a Resource Within a Dataset

The second step in the process involves creating resources within each dataset. The process includes:

  1. Locating the file
  2. Naming each resource
  3. Providing a description for the resource
  4. Specifying the file type

We also recorded a video showing how to add a resource for a dataset.

alt text "Add a resource")

Adding a Data View

  1. Navigate to the resource for which you wish to create the view
  2. Click on the "Manage" button
  3. Click on the "Views" tab
  4. Click on the "New View" dropdown and select an option. (Data Table is the preferred tabular data viewer. Data Explorer is outdated.)
  5. Name and configure the view.
  6. Click the "Add" button at the bottom to save the view of the data.

Multiple views are permitted. To make a particular view show up on the dataset landing page, navigate to that view and click the "Canonical View" button.

Add Data to Additional Groups

All data published through the open data portal should be assigned to one or more groups. Groups are topical categories that make it easier for users to find information by topical area. On the open data portal, there are 16 groups designated. The primary group assignment for a dataset should be made through the metadata record, however datasets can be assigned to multiple groups if appropriate.

Assigning datasets to additional groups can be done through the groups tab on the dataset "landing page".

Data Center Privacy Review: Publish New Datasets as Private

The Regional Data Center takes privacy seriously. We require all publishers to understand the privacy implications of data shared through the Regional Data Center.

More information on handling sensitive data is available in the privacy section of this publisher's guide. As one final check, we also require all datasets being published for the first time be marked as "private." Datasets can be published privately through the metadata.

The Regional Data Center will check to see if any datasets are ready for review prior to their going "live" at least once per weekday. If any potential issues are found with the data, you will be contacted by the Regional Data Center. If none are founnd, then Regional Data Center staff will publish your dataset as open data! We encourage publishers to proactively contact us by phone or email if there are any questions about privacy when publishing a dataset.

When updating resources within a dataset, there is no need to designate the dataset as private unless substantial changes affecting the design or composition of resources within the dataset have been made.


I uploaded a CSV or Excel file, but the resource page is not showing me the nice spreadsheet view of the data that I see on other pages. What might be wrong?

There's something in your file that is preventing it from being imported into our Datastore database. The three most common problems are:

  1. Unexpected characters (like spaces) in your column names: We recommend formatting column names in snake case. This means you should convert everything to lower case and change all spaces and punctuation to underscores (so "FIELD NAME" becomes "field_name" and "# of pirates" should be changed to "number_of_pirates").
  2. Unexpected changes in data type: If everything in a column is a number, but suddenly a word appears, that can trip up the import process.
  3. Non-ASCII characters or weird character encoding: These will often manifest as weird symbols when you view your file in a text editor.

Editing or Updating a Dataset or Resource

We recorded several additional videos showing how to edit or delete an existing dataset. Links to each video are included below.

Edit a Dataset
Delete a Dataset
Add a Resource to an Existing Dataset
Change the File in an Existing Resource
Reorder a Resource
Delete a Resource

Setting up an ETL job

ETL (Extract-Transform-Load) jobs are automated processes we create to ingest data from an external source (such as an FTP server or a web site) and convert it into records in a dataset on our CKAN data portal.

This is the typical workflow for creating an ETL job:

  1. Using the instructions above, create a dataset (CKAN calls it a "package"), keeping it in "Private" mode.
  2. Fill out as much of the metadata as you can, but definitely add a title, description, tags, license, the frequency of data change and publishing, and the data steward name and e-mail address.
  3. Upload each of the files that should be in the dataset (these files will eventually be overwritten by ETL processes, but uploading the files lets you set the names and formats of the resources, and provide descriptions if warranted).
  4. Read Creating a data dictionary and then create a data dictionary. For datasets with CSV files (or other tabular data), you don't need to upload a separate data dictionary file now that we have the integrated data dictionaries described in that link.
  5. Let our Data Magician know where to find the data (e.g., the file named "awesome_dragons_and_bathtubs.csv" on the FTP server), whether it's an incremental update (e.g., just the last month) or a dump of the entire history, and when the script is scheduled to push new data (e.g., every Monday at 2am). If there's a good primary key for the data table, let us know!
  6. We'll set up the ETL job and let you know. When everyone is satisfied that the dataset is done, someone (you or us) will switch it from "Private" to "Public".

Extra credit: More documentation!

If you want to add even more documentation to your dataset, beyond the dataset description and the metadata, talk with us and check out our Data Guides and also the "Datasheets for Datasets" standard for ideas on what kind of documentation could be useful to users.


Coming soon!