Creators: How to Publish Data Sources & Schedule Extract Refreshes
As of 2021, the new data source creation and publishing is managed ONLY by Eliza Epstein, Sam Perkins and Maria Lugovets.
If you need a new data source created:
- Please review sections below What is a Published data source in Tableau? and How to Plan a Data Source
- Then submit a request via Submitting a Support Ticket | Splash Knowledge Base - option Data Warehouse Support.
The team will follow steps below to complete your request.
What is a Published data source in Tableau?
It is a unique data source you have created and saved/published to the Tableau Server. By publishing it you and others within your team can easily link the data to your workbooks. It cuts down on processing time and decreases backend costs by making the data source a daily extract that refreshes to our Tableau Server. Most importantly it allows us to:
Collaborate and share with others
Allow people in your organization to view, interact with, download, subscribe to, share, edit, and save published views, even if they do not use Tableau Desktop. Incorporate views into blog posts or websites.
Centralize data and database driver management
Create and publish data models that everyone can use. Centralized data management allows for sharing a single source for your Tableau data. All workbooks connected to the published data reflect updates to it.
Support Mobility
Access your data from a different computer or location, through a web browser or the Tableau Mobile iOS app. Sign in to your organization's Tableau Server from a private network offsite.
See more here Best Practices for Published Data Sources
How to Plan a Data Source
Before creating and publishing a new data source, you MUST do the following:
1. Ask yourself, does your data source fit the following criteria?
A. We are using these same relationships between tables repeatedly for multiple workbooks
B. It is a larger dataset, say 20,000+ rows
C. It contains multiple complex relationships
2. Read the following article: Plan the Data Source – Tableau
3. Check existing published data source list to ensure no duplication 6. Published Data Sources - Tableau Online
4. Ask yourself, do I need to publish this data source?
You need to publish a data source only if you or your colleagues plan to use it and re-use data from it on regular basis (for example, a dashboard for a monthly report). If you just need to explore data in the data warehouse, use one of the existing Published data sources or use direct link to Snowflake (see How to access the data warehouse – Asana)
5. Think ahead! Best practice is to plan what data you will need/will not need based on a specific project/use case/dashboard you are planning.
EX) Will I need Site Details but no Infrastructure related fields? Will I need Population Data? Will I need some Commcare data too? If yes, from what Commcare table in the DWH? Is my data limited to only portion of Splash data, for example Project WISE sites only? Answers to these questions will help you determine what your data source is BEFORE publishing)
6. Use Salesforce Site ID to link 2 or more tables! 'Salesforce Site ID' is the most unique and relevant data point that is present in all the Splash data warehouse tables (SF Site ID example: 0016100000ammetAAA). Using other data points for linking tables will result in incorrect data aggregation and views.
7. Optimize your data source by cutting down on size and background data processing! Ways to do this include:
A. Create a data source in a blank workbook.
B. Apply filters to the data source based on the project needs (for example, exclude data for Site/Shipping Country other than India and Ethiopia).
C. Do not include tables or fields if you’re not using them.
D. Use Relationships between tables instead of Joins (default set up is 'Relationships' in Tableau).
How to Publish a data Source in Tableau
Once you created the desired new data source in a blank workbook and applied Data Source Filters (if applicable), continue to the steps below to make the new data source available for other Splash users.
The following steps give an overview of the publishing workflow you will use, regardless of the type of data or the server you publish to.
1. Select Server > Publish Data Source.
If your workbook is connected to multiple data sources, select the one you want from the Publish Data Sources submenu.
2. If you’re not already signed into Tableau Server or Tableau Online, sign in now.
3. In the Publish Data Source dialog box reference the most up to date Splash Tableau Key & Outline file details & do the following:
A. For Project, select the project you want to publish to and enter the data source name. This can be found under 6. Published Data Sources -> Programs or Development.
B. For the Name, use the table name in either title case or abbreviated format so you can easily identify what is contained within. Display these names in the order they are in relations/joined in, with ‘+’ between then to indicate their separation.
If any filters are applied to the data source, add it into the data source name : *Filter: ....
Ex) SF Site Info No Pop Served + PD Affiliation
SF_Site Info and Most Recent Population Served + SF_Infra + Water Quality Reports * Filter: Ethiopia and India only; Project WISE Group Non-blank only
C. For Description, put your name or initials as the creator and list all of the tables and filters contained therein.
Ex) ML: contains SF data for 1. Site, MOU Information, Behavior Change Reports, Affiliated Sites, Grant Funded Sites, Grant Allocation, Sites Grantee, Record Type 2. Affiliation
or
ML: contains SF data for 1. Site, MOU Information, Behavior Change Reports, Grant Funded Sites, Grant Allocation, Sites Grantee, Record Type, Most Recent Populations Served Grouped 2. Campus, Water Infrastructure, Hygiene Infrastructure, Sanitation Infrastructure, Other Infrastructure, Site Filter & Water Filter 3. Water Quality Reports Filter** excludes any Sites/Shipping Country that is not Ethiopia and India; excludes any sites that have Project WISE Group field blank
D. For Permissions, accept the default project settings.
E. For Authentication, if you need to provide credentials to access your data, you can specify how authentication should be handled when the data is published to the server. - preferred Splash approach is 'Embedded Password'.

4. If applicable: if you are publishing file-based data that is on a Windows mapped drive or using images that will not be available from the server, select Include external files.
5. By default, during the publishing process, Tableau updates the workbook connection to use the new published data source. It also closes the local data source.
To continue using the local data source instead, clear the Update workbook to use the published data source check box.

6. Click Publish.
7. Your data source will open in your internet browser. From there, if you did make a Live connection, change it to Extract. Then wait until the status changes from 'Data is Creating Extracts' to 'Data is Extract'. Depending on the size of the data set, this may take several minutes.


How & why do I keep my Extract refreshed?
When you publish to the server, if you do not refresh it, it will remain static and never update. This means the data will remain unchanged and only reflect data from the date of the original upload. In some cases, if your data is not going to change because it is from an excel for instance, you do not need to set up the refresh.
In case the refresh is necessary, please follow these steps:
1. Create a refresh schedule: Go to your Data Source ->Click on tab ‘Extract Refreshes’ -> click ‘New Extract Refreshes’
2. Select the check box for the workbook or data source you want to refresh, and then select Actions > Refresh Extracts.
3. In the Refresh Extracts dialog, select Schedule a Refresh, and complete the following steps:
A. Refresh Type: Select Full Refresh (by default). A full refresh is performed by default.
B. Refresh Frequency: Please ONLY set up for Daily, Weekly or Monthly between 8-9AM PST as this is the frequency that the data warehouse is set to refresh (setting up different frequency will not provide better data but will cause unnecessary costs.) Setup the frequency for the extract refreshes to run.
C. When choosing the day, only do days that you are certain you WILL be using the data on. Otherwise, the extract refreshing charges Splash without need.

4. Let other Tableau users know about your wonderful new data source! Email other Tableau Users on your team with the details of the Published Data source such as what is contained within and how it can be used.
After Publishing your Data Source, Publish your Workbook
When working in Desktop, you want to publish the workbook as well so that it is on the server and is usable online! This ensures your work is saved as well as allows others to interact with it.
1. Open up the workbook in Tableau Desktop.
2. Go to one of the sheets.
3. Add the new extracted data source to the sheet, by selecting Data in the Menu bar and selecting ‘Add New Data Source’.
4. Select Tableau Server, choose the extract you just made, and hit ‘Connect’
5. Control click on the data source you just added and select ‘Replace Data Source’

6. Confirm the replacement and hit 'Ok'
7. Control click on the previous live data source and select ‘Close’

8. Re-hide the sheets.
9. Go to Server in the menu bar and select ‘Publish the Workbook to Tableau Online’
10. Re-save the Tableau Desktop workbook and close.
11. Let other Tableau users know about your wonderful new workbook if you like! Not only can these be helpful to others, but it also allows people to get inspired! Email other Tableau Users or persons on your team with the details of the workbook such as what is contained within and how it can be used.