Text analytics

There are four components in the TEXT ANALYTICS group of the components menu:

TA Column Chart and TA Bar Chart — chart the incidence of tagged terms from text data. These must be connected to a data source.
Text Explorer Output — displays an output from a Text Explorer project.
Text Auto Report Output — displays a Text Auto Report.

TA Column Chart

Used to display the incidence of tagged terms from text data. Works with Noun Chunks, Parts of Speech, and Danger Words output. Displays values along the vertical axis, ranked in descending order. A TA Column Chart takes one data source.

TA Bar Chart

Used to display the incidence of tagged terms from text data. Works with Noun Chunks, Parts of Speech, and Danger Words output. Displays values along the horizontal axis, ranked in descending order. A TA Bar Chart supports unlimited data sources.

Properties

Component Name — enter a label for the component (optional). The label will not be visible in the dashboard; it is useful for identifying components quickly while laying out your dashboard.
Chart Format — edit chart titles, subtitles and font.
Size — specify the component’s height and fill settings.
Chart Settings — edit various chart settings such as data sorting, smooth data gaps, data labels, label format, and decimal places shown.
Colours — there are 10 default colours pre-set in Dashboard Publisher. You can change any or all of these by clicking the arrow next to each colour.
Column or Bar — switch between a column or bar chart.
Text Analytics — includes:
- Data Point Limit — the number of terms shown in the chart: 5, 10, 20, 50, 100 or 200. The default is 10.
- Stop Words — any words you don’t want to show in your chart.
- Vs Tagged Comments (the default) or Vs All Comments — the base used to calculate the incidence of each term.
Horizontal Axis — configure the horizontal axis settings.
Vertical Axis — configure the vertical axis settings.
Legend — configure the legend settings.
Border and Background — add borders and/or background colours to the component.

Text Explorer Output

Displays an output from a Text Explorer project on the dashboard page. It doesn’t use a data source.

Under Settings, go to TE Result Output and click Select Output.
In the Select Output dialog, choose the Text Explorer project and the output you want to show.

Text Explorer Output doesn’t support dashboard filter mapping.

Text Auto Report Output

Displays a Text Auto Report on the dashboard page. It doesn’t use a data source. Under Settings:

Project — select the Text Auto Report project to display.
Filters Mapping — pair dashboard filters with the report’s filters (see below).
Use New UI — display the report with the new Text Auto Report interface.
Enable Ask Eve — on by default.

Map dashboard filters

Click Filters Mapping to open the Pair Filters dialog, then pair each dashboard filter with a filter in the Text Auto Report. Drag rows to reorder them, and click Save when you’re done.

Text Auto Report filters are categorical only. They can’t be paired with text, timeframe or period filters. Dashboard filters that can’t be paired are listed under Excluded Dashboard Filters.
Report filters that are mapped to a dashboard filter are hidden in the report, so viewers use the dashboard filter instead.
If the selected filter values don’t make a valid combination for the report, the report shows a “Filter Gone Rogue!” message.

Frequently asked questions

What text analytics outputs work with the TA charts?

TA Column Chart and TA Bar Chart work with Noun Chunks, Parts of Speech, and Danger Words output. The connected data source must contain these outputs for the charts to display data.

What is the Data Point Limit?

The Data Point Limit controls how many terms are displayed in the chart: 5, 10, 20, 50, 100 or 200. The default is 10, meaning the chart will show the top 10 most frequent terms ranked in descending order.

What’s the difference between Vs All Comments and Vs Tagged Comments?

This setting controls the base used to calculate the incidence of identified terms. Vs All Comments compares the volume of identified terms against the entire base of comments. Vs Tagged Comments (the default) compares only against comments that contain at least one identified term. Choose the option that best suits your analysis — All Comments gives a broader perspective, while Tagged Comments focuses on the subset of relevant comments.

What are Stop Words?

Stop Words are terms you want to exclude from the chart. If certain common or irrelevant words are appearing in your text analytics results, add them to the Stop Words list to filter them out of the visualization.

Can TA charts connect to data explorations?

No. TA Column Chart and TA Bar Chart connect to a data store only — they do not support data explorations.

How many data sources can I add?

A TA Column Chart takes one data source. A TA Bar Chart supports unlimited data sources. Text Explorer Output and Text Auto Report Output don’t use data sources.

Can I visualise Text Scoring output in a dashboard?

Only some of it. The TA charts work with the Parts of Speech, Danger Words and Noun Chunks variables in the data store created by the wrangle. They don’t chart numeric scores such as sentiment or stress scores. To show these outputs, add a TA Column Chart component to the dashboard page and connect it to the relevant variable. See Process text for the variables each Text Scoring service creates.