Press "Enter" to skip to content

Category: Visualization

Visualizing a Table

Published 2019-11-01 by Kevin Feasel

Cole Nussbaumer Knaflic takes a fairly simple table and creates eight visuals out of it:

STEP 1: When I encounter this table, I start reading and scanning down columns and across rows. In terms of specific observations, I might start by noticing that the majority of accounts are in Tiers B and C, while Tiers A and A+—though they don’t make up a huge number (or percentage) of accounts—do make up a meaningful amount of revenue. In terms of questions, I wonder if the tiers are in order: I would think A+ belongs above A and am confused that they don’t appear that way in the table (perhaps due to alphabetical sorting?).

This is a really nice practical exercise if you want to learn how to apply the right visuals to tell your story.

Comments closed

Plotting Three-Dimensional Linear Models

Published 2019-10-29 by Kevin Feasel

Sebastian Sauer shows a few techniques for visualizing linear models with two predictors:

Linear models are a standard way of predicting or explaining some data. Visualizing data is not only of didactical value but provides heuristical value too, as demonstrated by Anscombe’s Quartet.

Visualizing linear models in 2D is straightforward, but visualizing linear models with more than one predictor is much less so. The aim of this post is to demonstrate some ways do visualize linear models with more than one predictor, using popular R packages. We will focus on 3D examples, that is, two predictors.

I have a strong bias against 3D visuals because they tend to be so difficult to see clearly. There are times when they’re necessary, though.

Comments closed

Task-Based Effectiveness of Visualizations

Published 2019-10-25 by Kevin Feasel

Adrian Colyer summarizes an interesting IEEE paper:

So far this week we’ve seen how to create all sorts of fantastic interactive visualisations, and taken a look at what data analysts actually do when they do ‘exploratory data analysis.’

To round off the week today’s choice is a recent paper on an age-old topic: what visualisation should I use?

No prizes for guessing “it depends!”

…the effectiveness of a visualization depends on several factors including task at the hand, and data attributes and datasets visualized.

Is this the paper to finally settle the age-old debate surrounding pie-charts??

The results were very interesting, though as an official Pie Chart Hater, I would point out that in none of their results was a pie chart ever better than a bar/column chart. There are cases where it works out okay, but if it’s never better and often worse than something, I’d rather use the alternative.

Comments closed

Head-to-Head Comparisons with Power BI

Published 2019-10-24 by Kevin Feasel

Rob Collie walks us through building a visual which provides head-to-head comparison using Power BI:

Yes, I know that NONE of the infographics above is a scientifically “good” comparison tool.  Too noisy, too flashy, not clean…  but every now and then you DO need to cater to your audience.  Engagement is the first step in the comprehension funnel, and in this particular example, yep, I’m trying to capture the eyeballs of an audience that likes this sort of thing.  The style of #4 is a decent compromise in this case.  Know your audience.

Rob takes us through an interesting journey. I don’t think I’d want to use that style too often, but to be fair, Rob talks about that in the snippet I clipped.

Comments closed

Key Performance Indicator Transparency

Published 2019-10-24 by Kevin Feasel

Treb Gatte explains how to build trust in KPIs:

Many Power BI reports and dashboards have simple, color-based indicators, commonly referred to as Key Performance Indicators or KPIs. These indicators provide a way to communicate the current condition of a monitored metric.

These KPIs require the data consumer to:
– Trust the data used to calculate it
– Understand the rule used to derive the indicator
– Understand the calculation used to set the value

Read on for good questions and a recommended design. Trust is critical, as people won’t look at dashboards whose data they don’t trust.

Comments closed

rBokeh Tips for Missing Arguments

Published 2019-10-23 by Kevin Feasel

Matthias Nistler walks through troubleshooting rBokeh missing argument errors:

This approach is my go-to solution to change a rBokeh plot for which there is an argument missing in rBokeh that is available in python.
– Create the plot.
– Inspect the structure (str(plot)) of the rBokeh object.
– Search for the python’s argument name.
– Overwrite the value with the desired option as derived from python’s bokeh.

Given how nice the bokeh package looks, I really want rBokeh to work well. Hopefully this experience improves over time.

Comments closed

Shaded Ranges in Excel

Published 2019-10-16 by Kevin Feasel

Elizabeth Ricks shows how to create shaded ranges in Excel:

We can see there’s clear seasonality in this business—overall volume is highest in the summer and each outing type generally follows the same monthly pattern. Let’s say you manage the Family rentals and you’d like to compare your monthly volume to what you’re seeing across the entire fleet. 

For the purpose of this tactical illustration, let’s assume the shape of the data—relative peaks and valleys—is more important than the specifics of each category individually. If that’s the case, I can simplify by showing a shaded region to depict the range of absolute passengers each month.

This technique is excellent when you have a large number of lines but only care about one versus the norm, and individual lines would be too distracting.

Comments closed

Fun with Residual Plots

Published 2019-09-30 by Kevin Feasel

Nina Zumel explains why, when plotting residuals, you always put predictions on the X axis and residuals on the Y axis:

One reason that the proper residual graph (for a well fit model) should smooth out to the line y=0 is known as reversion to mediocrity, or regression to the mean.

Imagine that you have an ideal process that always produces a single value y. You don’t actually observe this “true value”; instead, what you observe is y plus (IID, zero mean) noise. You can build a “model” for this process that predicts the mean of the observations, in this case the value 0.1033149. Then you can calculate the residuals of your “model” in the usual way.

This post went in a direction I wasn’t expecting, and it was all the better for it.

Comments closed

Storytelling with Power BI

Published 2019-09-30 by Kevin Feasel

Marc Lelijveld wraps up a series on storytelling with Power BI:

Let them ask questions
As a report author, you start building your reports based on the information needs and business requirements you collected before your project. However, every answer to a question, triggers a new question to come up. In the end you end-up with more questions to answer than you thought about up front. Maybe even with scope creep in agile projects.

However, it is very unlikely that you answer all the business information needs in your dashboard or report within one iteration. So why not give them the ability to exploitative interact with the report and ask questions in a native language to their dataset? Power BI has the ability to ask questions to your data in your native language in just a few clicks.

This is probably one of the most underutilized aspects of Power BI.

Comments closed