Press "Enter" to skip to content

Category: Visualization

Improving Diagrams

Published 2019-05-02 by Kevin Feasel

Stephanie Evergreen gives us some tips for improving diagrams:

As ubiquitous as diagrams are, they have some significant shortcomings.

Diagrams tend to be most useful as a mental organizing activity for the people who make the diagram. Outsiders coming in fresh have a much harder time seeing how the elements of the diagram fit together and make sense. I think that is generally because diagrams lack enough narrative to explain what is going on. We often use diagrams without explaining them. Or, more precisely, we use diagrams without connecting the diagram pieces to their associated parts of our narrative.

Read on for an example.

Comments closed

Controlling Power BI Visual Visibility

Published 2019-04-25 by Kevin Feasel

Matt Allington shows how we can take one Power BI visual and use it to control the visibility status of another visual:

I have written a few articles in the past that toy with the ideas of changing visibility and text colour based on selection.  I started to wonder if it was possible to make a visual appear (or not) based on a selection from the user.  There is no out of the box way to do that today. It is possible to use bookmarks to show an hide an object, but the user must click a specific button to do this. I want the user to be able to interact with a report and see (or not see) a chart based on some valid selection across the report.  Microsoft is already working on building expression based formatting across the breadth of Power BI however as of now the only item you can change is the header in a chart.

Hopefully this gets better over time.

Comments closed

Modifying HTML Rendering in Shiny

Published 2019-04-22 by Kevin Feasel

Senthil Thyagarajan gives us an example of uisng the htmltools package to change the way tables render in Shiny:

In order to build the html table I have used a function table_frame which can be used as a container in DT::renderdatatable. This function basically uses htmltools. For more references on the basics of html tables please refer here

In addition to changing the colors, Senthil also shows how to add a couple of buttons which call Javascript functions. H/T R-bloggers

Comments closed

Qualitative Analysis with Dendograms

Published 2019-04-19 by Kevin Feasel

Stephanie Evergreen explains what denodgrams are and when they can be useful visuals:

Dendrograms are not THE most common qualitative visual because they require a data generated through a hierarchical cluster analysis. Cluster analysis can be a useful tool in analyzing qualitative data. By clustering groups of participants with similar qualitative codes, you can better understand your findings. According to Henry & team, this analysis can help “reveal things like participant motive and the reasons behind counterintuitive findings.”

Check out Henry’s article to learn more about the analysis. Here, let’s just focus on describing a dendrogram that could display those hierarchical cluster analysis findings. They can be a little confusing at first, especially since the x-axis has 100% closest to the y-axis when we aren’t used to seeing it that way. Walk through this example with us.

Click through for an example. If it’s confusing at first, read to the end, as I think the concrete example helps everything click.

Comments closed

Telling a Story When Data is Always Changing

Published 2019-04-11 by Kevin Feasel

Meagan Longoria explains how you can tell a story with data when the data is not static:

If you are a BI/report developer, you know this challenge well. You may follow all the guidelines: choose a good color palette, make visuals that highlight the important data points, get rid of clutter. But what happens when your data refreshes tomorrow or next month or next year? It’s much easier to make a chart with static data. You can format it so it communicates exactly the right message. But out here in Automated Reporting Land, that is not the end of our duties. We have to make some effort to accommodate future data values.

Meagan uses the Phone-a-Friend option and gets an interesting inversion of the normal solution.

Comments closed

Finding Dependency Clusters

Published 2019-04-04 by Kevin Feasel

Michael J. Swart performs cluster analysis with tables:

That ball of mush in the middle is hard to look at, but the smaller disconnected bits aren’t! Just like Ben, I want to work on those smaller pieces too! And just like the lonely tables we looked at last week, these small isolated components are also good candidates for extracting from SQL Server.

The script looks at joins in execution plans, which is a rather clever way of doing this when you don’t have a comprehensive set of foreign key constraints.

Comments closed

Shrinking Dot Sizes in Power BI

Published 2019-04-02 by Kevin Feasel

David Eldersveld shows how we can reduce the point size of dots in POwer BI as of March 2019:

One of the Power BI improvements in the March 2019 Desktop release was reduced bubble size for the Map visual. I previously wrote about the benefit of the reduction in point/bubble size. I was unaware until recently that this change made it into more than the Map visual.

The ability to reduce the point size also appears in the Format options for the Power BI Scatter chart. Previously, you could change the size option from 0 to 100 under the Shapes area. As with the Map, the Scatter now allows you to reduce the size as low as -30. I did not see this mentioned in the March Desktop blog post. I must have missed it if it was part of a previous month’s release. In any case, if you were not aware that you could set the point size from -30 to 100with the Scatter chart, now you do.

For most scenarios, I think the dot size is probably a little too big. -30 is generally too small, but I’m happy that they offer us options to get it right.

Comments closed

Databricks Dashboards

Published 2019-03-27 by Kevin Feasel

Megan Quinn takes us through building dashboards with Apache Zeppelin on Databricks:

The first step in any type of analysis is to understand the dataset itself. A Databricks dashboard can provide a concise format in which to present relevant information about the data to clients, as well as a quick reference for analysts when returning to a project.

To create this dashboard, a user can simply switch to Dashboard view instead of Code view under the View tab. The user can either click on an existing dashboard or create a new one. Creating a new dashboard will automatically display any of the visualizations present in the notebook. Customization of the dashboard is easily achieved by clicking on the chart icon in the top right corner of the desired command cells to add new elements.

This isn’t quite a step-by-step guide but does spur on ideas.

Comments closed

Conditional Formatting on Text Fields in Power BI

Published 2019-03-26 by Kevin Feasel

Matt Allington shows how you can apply conditional formatting to non-numeric fields in Power BI:

The high level process is to:
1. Create a measure that returns a colour as the result

1. It can be a word, such as blue, red, green
2. It can be a hex code for a colour, like #40E0D0″, “#FFA07A”
2. Use conditional formatting and use the measure to apply the formatting on the text as a rule.

Read on for a demo.

Comments closed

Running Totals in Tableau and Power BI

Published 2019-03-25 by Kevin Feasel

David Eldersveld shows how to create running totals in both Tableau and Power BI:

What about a separate Power BI Date table?
This setup is built for consistency of comparison. As people go deeper into Power BI, they typically add a separate Date table as part of a more robust data model and add relationships between tables. At the same time, they disable the default Auto Date/Time built-in hierarchies. This more advanced setup with a separate Date table allows several conveniences as well as performance and storage benefits. It’s especially true with larger models that include many facttables that each join to Date and other possible dimension tables. Tableau doesn’t currently have a comparable data model. We’ll stay conveniently away from that setup in Power BI because we only have one simple sample table.

I think both of them make this an easy operation, though Tableau is probably easier here.

Comments closed