Hierarchical Clustering for Rows

Row-based hierarchical clustering groups observations, such as customers, products, or survey respondents, based on their characteristics.

For example, you can group customers using their purchase frequency, spending, and product preferences to identify meaningful customer segments.

Exploratory displays the results through an interactive dendrogram, allowing you to explore how observations are related.

You can also change the number of clusters interactively to explore different levels of detail.

For instance, you might start with three clusters, examine their characteristics, and then increase the number to four or five to discover more specific segments.

Hierarchical Clustering for Columns

Exploratory also supports hierarchical clustering for columns (variables).

This is especially useful when analyzing datasets with many related variables, such as survey questions or product evaluation criteria.

For example, if a survey asks respondents to rate products on price, quality, convenience, design, and brand trust, hierarchical clustering can help identify which questions show similar response patterns.

This makes it easier to discover groups of related variables and understand the relationships among them.

With hierarchical clustering joining existing methods such as K-Means, K-Modes, and K-Medoids, Exploratory now offers even more ways to discover meaningful patterns in your data.

Calculate NPS Directly in Pivot Tables

Net Promoter Score (NPS) is a widely used metric for measuring customer loyalty and willingness to recommend a product or service.

It is typically calculated from responses on a 0–10 scale:

• Promoters: Scores of 9–10
• Passives: Scores of 7–8
• Detractors: Scores of 0–6

NPS is calculated as:

NPS = % Promoters − % Detractors

Previously, calculating NPS required additional steps to categorize respondents and calculate the percentages.

In v16.4, you can now calculate NPS directly from the Pivot Table menu.

This makes it easy to compare NPS across customer segments, regions, age groups, or other categories without manually preparing the calculations.

Compare Overall Results with Individual Groups

Exploratory’s Repeat feature lets you create separate charts or tables for different groups, such as age groups, regions, or customer segments.

With v16.4, you can now include the overall results alongside the individual groups.

For example, when comparing customer satisfaction across age groups, you can display the overall distribution next to the distributions for customers in their 20s, 30s, and 40s.

This makes it much easier to understand how each group differs from the overall population without creating separate charts or tables.

The enhancement is not just for Pivot Tables, but also for other chart types like bar chart.

And you can even have the total bar inside the bar chart. Just select ‘Show Total Bar’ at the X-Axis menu.

Display Statistical Test Results on Box Plots

Box plots are useful for comparing data distributions across groups, but visual differences don’t necessarily indicate statistically significant differences.

In v16.4, you can now display statistical test results directly on box plots.

For example, you can show p-values from t-tests or Wilcoxon tests alongside the distributions you’re comparing.

This lets you examine both the distributions and statistical evidence in a single visualization, making your analysis easier to interpret and share.

Customize Marker Shapes in Scatter Plots

Scatter plots now support assigning a variable to Shape, allowing different categories to use distinct markers such as circles, triangles, and diamonds.

For example, when exploring the relationship between customer spending and purchase frequency, you can distinguish customer segments using different marker shapes.

Combined with existing color and size options, this provides another way to visualize multiple dimensions within a single chart.

Customize Line Chart Markers

Line charts now offer additional options to customize markers and shapes for individual lines.

This helps distinguish multiple series more clearly, especially when charts are printed in black and white or when colors alone aren’t sufficient.

Improved Trendline Selection and Tooltips

We’ve redesigned the trendline selection menu to make it easier to choose the right model.

Trendline types are now organized into structured categories, with explanatory tooltips to help you understand their purposes.

We’ve also improved the trendline hover tooltip.

Information associated with the cursor position, such as predicted values and confidence intervals, is now clearly separated from model-level statistics such as R², coefficients, p-values, and observation counts, where applicable.

This makes it easier to understand both the model’s predictions and its overall statistical characteristics.

Edit Dashboards Directly in Preview Mode

You can now edit dashboards directly in Preview mode.

Instead of repeatedly switching between editing and previewing, you can make adjustments while seeing how your dashboard will appear.

This streamlines the process of refining layouts, arranging charts, and making design changes.

Add Images to Dashboards

Dashboards now support image components, allowing you to include company logos, product images, illustrations, and other visual elements.

You can also use parameters within image URLs to dynamically change displayed images based on user selections.

For example, selecting a product from a dashboard parameter can automatically update the product image.

More Flexible Dashboard Navigation

Dashboard page tabs can now be positioned on the left or right side, in addition to the top.

Vertical navigation is particularly useful for dashboards with many pages or long page titles.

We’ve also added support for displaying tabs across multiple rows, helping keep larger dashboards organized and easy to navigate.

Expanded Font Customization

You can now customize font families throughout your dashboards, with improved controls for titles and other text elements.

These enhancements help maintain consistent typography between Exploratory Desktop and published dashboards.

Choose Which Pages to Publish

When publishing a dashboard, you can now choose to hide specific pages or tabs.

For example, you can keep work-in-progress or internal analysis pages out of the published navigation while sharing completed pages with your audience.

Search Across Projects

A new search dialog on the project list page lets you find projects using keywords and structured filters.

This makes navigating large collections of projects much easier.

Search Within Projects

You can now search across a wide range of content within a project, including:

• Data wrangling steps
• Notes and their contents
• Scripts
• Charts and other project components

Search results include highlighted keyword matches and contextual snippets, helping you identify exactly where a term appears.

You can also jump directly from a result to the corresponding chart, step, or note.

For example, if you remember creating a calculation for customer retention but don’t recall where it was, you can search for the relevant keywords and quickly locate the step.

With improved indexing and relevance ranking, finding and reusing previous analytical work is now much easier.

Export Step Diagrams

Exploratory automatically records data wrangling operations as a sequence of steps, making it possible to review how data has been transformed.

The Step Diagram visualizes these operations and their relationships.

In v16.4, you can now export Step Diagrams as images or files.

This is especially useful when documenting analytical workflows, explaining data transformations to colleagues, or sharing how a dataset was prepared.

Improved Step Diagram Layout and Zoom

We’ve improved the positioning of boxes and connecting arrows in Step Diagrams, particularly around join operations and workflows involving multiple data frames.

The initial zoom behavior has also been enhanced to display more surrounding steps when opening a diagram.

This makes it easier to understand the broader workflow without manually zooming out.

Start Working with Large Datasets Faster

When importing a dataset, you can now choose whether Exploratory initially displays the Summary View or Table View.

For large datasets, selecting Table View allows you to start inspecting and working with the data without waiting for summary statistics to be calculated.

This can make a noticeable difference when working with datasets containing many rows or columns.

Swap Chart Variables with Drag and Drop

You can now drag and drop variables between chart shelves, including X Axis, Y Axis, Color, and Repeat.

This makes it faster to explore different visualizations without repeatedly selecting variables.

Move Repeat Labels to the Top

You can now position Repeat labels above individual charts rather than below them.

This improves readability when comparing multiple charts side by side.

Enhanced Show Details Dialog

The Show Details dialog now supports resizing column widths through drag and drop.

You can also select which columns to display, making it easier to focus on relevant information and keep unnecessary columns out of the view.