Preparing Data
Building a dashboard doesn’t start in Dashboard Publisher. Dashboard components can only connect to an existing data source, so you’ll need to prepare your data in other applications such as Survey Shaper, Data Explorer, and/or Data Wrangler.
Think of this as your pre-flight checklist. Before you dive into building, here’s what you’ll need to consider and prepare:
Where’s your data coming from?
Section titled “Where’s your data coming from?”First things first—identify your data source. Will you be working with Survey Shaper data or an external Data Store?
Whichever it is, the data store must be published before Dashboard Publisher can use it. Data stores filled by importing a file are published automatically when the import finishes, and so are survey data stores and the data stores wrangles create. A data store you build variable by variable stays in Draft until you publish it. See Publish & manage a data store.
What variables do you need?
Section titled “What variables do you need?”- Make sure your survey or external data source has all required variables in the correct format
- We recommend sketching out a mockup of how you want to present the data on your dashboard—this helps you visualize what you need
What filters will you use?
Section titled “What filters will you use?”Make sure you have a variable in your data source for each filter variable you plan to use.
How will you structure your data?
Section titled “How will you structure your data?”- Data visualization components in Dashboard Publisher can connect to either a Data Store or Data Exploration
- Filters and Verbatim questions connect directly to the Data Store, but for all other variables, you’ll want to connect to Data Explorations because they offer much greater flexibility
- In Data Explorer, you’ll create an exploration for each variable you want to put in the dashboard (aside from the filter and verbatim variables)
Will any pages need low base blocking?
Section titled “Will any pages need low base blocking?”If a page shouldn’t show results when too few people have responded, you can turn on low base blocking for that page. The check reads its base size from a data exploration, so create an exploration that shows the base (count) you want to check. When you connect it, you’ll select the one cell that holds the base size. See Low base blocking.
Creating Dummy Data
Section titled “Creating Dummy Data”If the data source is a Survey Shaper survey where data collection hasn’t started yet, consider creating dummy data before opening Data Explorer. This gives you something to work with while you build your dashboard.
- Publish the survey in Survey Shaper
- Copy the live published survey link and paste it into your browser
- Complete a few surveys, making sure you’ve provided a response to all questions you want to put in the dashboard
Next Steps
Section titled “Next Steps”Once your data is prepared, you’re ready to create your dashboard layout.
Frequently asked questions
Section titled “Frequently asked questions”What types of data sources can I connect to dashboard components?
Dashboard Publisher supports three types of data sources:
- Data Store — connect directly to a data store and select individual variables with an aggregation type (Mean, NPS, CSat, Percentage, Categorical, Multi-Response or Comment)
- Data Exploration — connect to a saved exploration from Data Explorer, which provides pre-configured cross-tabulations
- Manual Query — enter a direct query. This is an advanced feature that isn’t available to most users. Contact your administrator if you need it.
Most chart components support both Data Store and Data Exploration connections. Some components — like Comments — only support Data Store connections. Heading and Paragraph components have no aggregation types, so they can only be connected through Manual Query.
What variable types are supported in data stores?
Dashboard components can use the following variable types from data stores:
| Variable type | Description |
|---|---|
| Number (Decimal) | Numeric values with decimals |
| Integer (Numeric) | Whole number values |
| Boolean | True/false values |
| Text (String) | Short text values |
| Multi-Response | Multi-select response data |
Long Text, Date, DateTime and Time variables can’t be connected to dashboard components. Date and DateTime variables are used only by timeframe filters.
Not every variable type is compatible with every aggregation. For example, Mean only works with Number, Boolean, and Integer types, while Comment only works with Text types. Dashboard Publisher automatically filters the available aggregations based on the variable type you select.
Which aggregation types are available for data store connections?
The available aggregation types depend on your variable type and the component you’re configuring:
| Aggregation | Description | Compatible variable types |
|---|---|---|
| Mean | Average of the chosen variable | Number, Boolean, Integer |
| NPS | Net Promoter Score (groups values into Detractors 0–6, Passives 7–8, Promoters 9–10) | Number, Integer |
| CSat | Customer Satisfaction (groups values into Low 0–4, Mid 5–7, High 8–10) | Number, Integer |
| Percentage | Displays your chosen variable in percent format, based on the values you choose to include. Available for Single Value Tile and Smile Badge Result only | Number, Boolean, Integer |
| Categorical | Displays response categories | Boolean, Integer |
| Multi-Response | Displays multi-select response data | Multi-Response |
| Comment | Displays text/verbatim responses | Text |
Some component types only support a subset of these aggregations.
What are the requirements for data explorations used in dashboards?
Data explorations connected to dashboard components can have at most one row variable and one column variable. Explorations with stacked or bannered variables are only supported by the Data Exploration component.
If you connect a stacked or bannered exploration to any other component, you’ll see an Invalid Exploration message: “This exploration features Stacking and/or Bannering which are only supported when using the Data Exploration component”. To fix this, simplify your exploration in Data Explorer so it has no more than one variable in each dimension, then connect it to your dashboard component again.
How many data sources can a single component have?
It depends on the component type:
- Column, Bar, Line and Area charts (including the stacked and spline versions) and TA Bar Chart — unlimited data sources. You can overlay multiple data series on the same chart.
- Pie, Donut, Semi Circle and Item Chart — one data source only
- TA Column Chart — one data source only
- Comments component — one data source only
- Info Tiles (Single Value, Smile Badge) — one data source only
- Text components (Heading, Paragraph) — these can only be connected through Manual Query, which is an advanced feature that isn’t available to most users
- Containers — none (containers are layout elements only and don’t connect to data)
The Scatter Plot (up to 2) and Distribution (up to 4) components are also advanced features that aren’t available to most users.
When a component supports multiple data sources, each data source can connect to a different data store or exploration.
What happens if a data exploration is deleted after I’ve connected it to a dashboard?
If a data exploration is deleted from Data Explorer after being connected to a dashboard component, Dashboard Publisher detects the missing exploration when you next open the dashboard in the editor, and displays an Unable To Find Data Exploration alert:
“It looks like one of your Data Explorations has been moved or deleted. Please check your Workspaces.”
The affected component will show an error state. To fix it:
- Open the dashboard in Dashboard Publisher.
- Click on the affected component.
- On the DATA SOURCE tab, remove the missing data source.
- Add a new data source pointing to a valid exploration.
A missing exploration doesn’t stop you publishing, but the affected component won’t show data until you fix it, so resolve all missing data sources before you publish.
Do I need to create separate explorations for each chart on my dashboard?
Generally, yes. Since each data exploration can have at most one row variable and one column variable, you’ll typically need a separate exploration for each metric you want to visualize. For example, if you want to display three different survey questions as separate charts, you’ll need three explorations.
However, if multiple charts display the same data with different visual presentations, they can share the same exploration.
Tip: Organize your explorations in a dedicated folder within your Data Explorer workspace so they’re easy to find when connecting components.
Can I use code frame variables in my dashboard?
Yes. A variable’s code frame is used to label its values on your dashboard. When you connect a data store variable, choose the aggregation that suits the component:
- Categorical shows every code from your selected question.
- Percentage and Multi-Response prompt you to choose which coded values to include.
Code frames are also used to list the options in dashboard filters. For more control over how coded values are grouped and displayed, connect the component to a data exploration instead.
What’s the difference between connecting to a data store vs. a data exploration?
| Data Store | Data Exploration | |
|---|---|---|
| Best for | Filters, verbatim comments | Charts and metrics |
| Flexibility | Basic aggregations (Mean, NPS, etc.) | Full cross-tabulation capabilities |
| Configuration | Select variable + aggregation type | Select a pre-built exploration |
| Updates | Reflects latest data automatically | Reflects latest data automatically |
We recommend using Data Explorations for most chart components because they offer greater flexibility for data analysis. Use Data Store connections for dashboard filters and Comments components, which require direct variable access.
Do changes to a data exploration automatically update the dashboard?
New responses in the underlying data store flow through to the dashboard automatically. If you change the exploration itself, preview the dashboard to check the component still displays correctly, then republish to make sure viewers see the change. If you change the structure of the exploration (e.g., swap out variables), make sure it still has no more than one row variable and one column variable, or the dashboard component may show an error.
Why can’t I find a variable when I connect a component or add a filter?
Fields that have been excluded from exports and AI in the data store aren’t available in Dashboard Publisher. They don’t appear in the variable lists when you add a data source, a filter or a timeframe filter.
If a component or filter was already connected to a field before it was excluded, it shows an error icon instead of data in the dashboard. Remove that data source or filter, or connect it to a different variable. This also applies to Focus Words components whose rating variable or mapped filters use an excluded field.
If you need a field that has been excluded, contact the data store owner or your administrator.