Professional Documents
Culture Documents
(You can tab through ‘em real fast if you don’t care).
This is an interactive Power BI report that tell you what Custom
Visuals for Power BI are available. (It’s a fun way to see what a web-
embedded Power BI report looks like, too)
This is the (near) future….it’s still in preview mode (i.e. not widely available).
Meanwhile, there’s 33 years’ worth of charting hacks available freely on the web….
…most notably from this guy:
https://peltiertech.com/
Jon Peltier,
Jon Peltier,
PatronExcel
Saint of Excel
Charting Charts
Guru
Jon’s blog shows how to coerce
pretty much any “respectable”
chart from Excel
…and he sells an excellent charting add-in
https://peltiertech.com/Utility30/
Limitations can be liberating
Me:
So what’s the difference to you between Excel and Power BI?
When you open Power BI, you know exactly what you are meant to be doing:
“Business Intelligence”
Hirers are starting to explicitly call out that when it comes to BI Dashboards,
user-centered visual design skills are as important as technical IT skills.
“At
Visual the heart(1997)
Explanations of quantitative
covers Information reasoningof mechanism and
the representation
The Visual Display of Quantitative (1983) A highly regarded book
is
motion, a
Envisioningsingle
process
Beautiful
on statistical question…
and
Information
graphics, dynamics,
(1990)
Evidence
charts, causes
showcases
(2006)
tables. and
isTheory effects,
aboutsuperb
and explanation
displays
the theory
practice inofthe
and and narrative,
high-dimensional
practice
designof of
analytical
data design.
and design
complex
graphics, strategies
data
The
250 and for enhancing
bookprovides
identifies
illustrations bestthe
ofpractical
excellent
the rate
advice
and
(and of information
about ofhow
aeffective
few to
methods
the worst)transfer
explain in print,
for complex
showing
statistical nearly every
presentations,
material
graphics, by visual
kind
with and
of computer
means,
detailed withscreens.
information,
analysis extraordinary
suggests
of how tomany examples
newdata
display to illustrate
designs
for (including
precise, the
sparklines)
effective,
fundamental
quick analysis.principles of information displays.
Better comparisons = Better conclusions
We look at one page at a time. The more on the
page, the more effective and comparative our eye
“At the heart of quantitative reasoning can be.
is a single question…
Graphics can be shrunk way down. But as visuals shrink, the
clutter of chartjunk, non-data-ink, and redundant data-ink is
even more costly than usual in data-rich designs.
what?” when it introduces new means of engaging with data that simply could not
be achieved in any other way.”
Stephen Few: What graphs work best for dashboards?
“Compared to
what?”
Stephen Few’s “Approved” dashboard graphs
The best graphs for displaying quantitative information on a dashboard are:
• Bullet graphs Just because it is on this list does not make it a good choice!
http://www.perceptualedge.com/articles/misc/Bullet_Graph_Design_Spec.pdf
Only use a stacked bar chart when you must display multiple instance of a
• Bar charts whole and its parts, where the emphasis on the whole and comparisons of
• Dot plots Remember
the constituent Stephen Few’s
parts are less Principle
important.#1:
• Line charts It must be the best way to display the information.
Why? Difference are difficult to detect for all but the series that appears at
• Sparklines the bottom of each bar.
• Box plots If something else on the list makes it easier for the end user to answer the
question
Let’s “Compared
take a look at how atobad
what?” then
use of use that
a stacked bargraph
chartinstead.
can be rectified…
• Scatter plots
https://peltiertech.com/bullet-charts-in-excel/
• Spatial maps
• Heat maps Let’s run through a few of these quickly.
• Tree maps
“Compared to
what?”
Stacked Bar Chart Alternatives Much better way to arrange the same data
https://chandoo.org/wp/incell-panel-chart/
https://peltiertech.com/some-comments-on-dot-plots-guest-post/
https://peltiertech.com/stacked-bar-chart-alternatives/ for healthier consumption…
“Compared to
what?”
Stephen Few’s “Approved” dashboard graphs
The best graphs for displaying quantitative information on a dashboard are:
• Bullet graphs
• Bar charts
https://peltiertech.com/dot-plots-microsoft-excel/
• Dot plots
• Line charts As with Stacked Bar Charts, keep the number of series displayed in any
• Sparklines Stacked Line or Stacked Area charts “digestible”.
• Box plots Let’s take a look at how a bad use of a stacked bar chart can be rectified…
• Scatter plots https://peltiertech.com/dot-plots-microsoft-excel/
• Spatial maps
• Heat maps
• Tree maps
“Compared to
what?”
My redesign is a much better way to arrange
the same data for healthier consumption…
(Otherwise the user is forced to constantly move I use an amended version of Jon Peltier’s Label Last
their eyes between the series and the legend in Point routine.
order to work out what each series is. And that Note the simple
(I’ve amended the codepies
so itshowing
also colours the share
the labels thatthe
to match each group
series, and of
gets in the way of the “compared to what?”) products
auto-spacescontributes
the labels so theytodon’t
theoverlap.
total.ItHelps withthe
even redraws thelabels in real
time if people filter the chart to show/hide other series. I’ll be releasing it as a
‘Compared to what?’ question between graphs.
commercial add-in for Excel…ask me if you want to buy a copy)
Andy Kirk: “It is vital to treat every representation challenge on its own merits. Just
because you have spatial data does not mean that the most useful portrayal of that
data will be via a map.”
Lessons from a dashboard competition
Perceptual Edge 2012 Dashboard Competition: http://www.stephen-few.com/blog/page/16/
Stephen Few discussing Jason Lockwood’s winning entry (produced using Photoshop):
One of the first things you probably notice is its fine aesthetics. Its use of color, layout, and reduction of non-data-ink make it pleasing to the eye in a way that
enhances usability. Because color has been used sparingly, the red alert icons make it easy to spot the students that are most in need of immediate attention
(although the icons could be a little bigger to make them pop more).
The tabular (rows and columns) arrangement of student information (one student per row) makes it easy to see everything about a particular student with a
quick horizontal scan and easy to compare a particular metric across all students with a quick vertical scan.
All of the most important metrics were consistently represented using the same dark shade of blue, which featured them above other items nicely (although
the dark blue horizontal bars in the bullet graphs would have been easier to see and compare if they were thicker).
This design is scalable in that the addition of more students could be easily accommodated by simply expanding the dashboard vertically. Meaningful patterns
in individual student attendance information (days absent and tardy) can be easily seen.
“Compared to
what?”
Lessons from a dashboard competition
Perceptual Edge 2012 Dashboard Competition: http://www.stephen-few.com/blog/page/16/
Stephen Few discussing runner up Shamik Sharma’s dashboard (produced using Excel 2010):
Once again, notice the fine visual aesthetics of this design. Also notice the additional class-level information that appears on the right that doesn’t appear in
the winning dashboard, especially the two distribution graphs on top, which are quite useful.
And finally, notice how the frequency distribution graph of assessment scores in the bottom right corner, which uses lines (called a frequency polygon) is easier
to read than the one that uses bars (called a histogram) in the winning solution.
A few features in this solution don’t work as well, however, as those in the winning solution. For example, it isn’t as easy to spot the students in need of
attention, and per student attendance information is aggregated in a way that hides patterns of change through time. Overall, however, this is excellent work.
“Compared to
what?”
Lessons from a dashboard competition
Perceptual Edge 2012 Dashboard Competition: http://www.stephen-few.com/blog/page/16/
Writeup:
https://www.perceptualedge.com/blog/?p=14
66
It is rendered in Microsoft
Excel, but distributed via
Let’s
automatic look
email as aat a repor
PDF,
with the help of someto
Compared pretty
what?
trivial VBA code (Macros).
Andy Kirk:
Approx. 5% of people have compromised ability to discern particular colors/combos
such as Deuteronopia (red-green color blindness).
=> If your audience is undefined you are going to alienate potentially 1 person in
every 14!
ACTUAL DATA:
This future-focussed heatmap shows when staff are likely to be fatigued, but not why they
were fatigued in the first place. Plus it only covered 13 days, even though shifts planned 28
days in advance. Furthermore it only showed people that were over the reporting
threshold today but not who might be fatigued in the coming 28 days.
SELF ASSESSMENT TOOL
For instance, using actual charts means we can programmatically colour each individual bar of the mini-chart
‘sparklines’ to better draw out to users when fatigue levels are unacceptably high.
And we can also show multiple series in each chart like I’m doing here, instead of just one series. (In this case, coloured
bars for days worked, white bars for days off, and a dotted reference line showing the fatigue reporting threshold.)
We also moved from Excel
2010 to Excel 2016, which
gave us a LOT more options in
the ‘back end’ behind the
scenes…
SELF ASSESSMENT TOOL
The original Excel 2010 file had something like 1.3M cells
with formulas in them (618k with computationally expensive
VLOOKUPS ).
Everyone Safe
Every Day
In Excel 2013 or later, this complexity can be
replaced with the ‘Data Model”: Tables linked
by relationships between them
“Modern Excel” (and Power BI) : Hello Data Model, goodbye formulas…
With any version of Excel 2013 or later, you can use the Data Model to
mash together related tables directly to a PivotTable, without VLOOKUP.
These new PivotTables let you write formulas (called Measures) directly
inside Values fields using the DAX (short for Data Expressions) formula
language.
DAX is somewhat similar to traditional Excel formulas, and gives you
incredible control over how data is filtered, displayed, and aggregated.
Premium versions (or any Office 365 subscription plan) also include
PowerPivot, which additionally lets you create Calculated Columns that
give you similar control over fields it the other areas of your PivotTables.
https://support.office.com/en-us/article/where-is-power-pivot-aa64e217-4b6e-410b-8337-
20b87e1c2a4b
=> PowerPivot is the User Interface (UI) to the Data Model, but you can use DAX without
it. http://dailydoseofexcel.com/archives/2017/07/10/look-ma-no-powerpivot/
Matt Allington’s book and video course is the easiest way for Excel folk to learn Data
Modelling and the DAX formula language
See https://exceleratorbi.com.au/courses-overview/
“Modern Excel” (and Power BI) : Hello PowerQuery, goodbye cutting & pasting…
Similarly, with any version of Excel 2016 or later, a new inbuilt tool called PowerQuery
makes it much easier to do all your data import/transformations.
Ken says: “I love this tool. Not because of what I can do with it, but because of how easy it
makes it to get things done for a business pro without the need for coding.”
Ken is coming to Wellington in February 2019, and will be running a paid workshop if I can find some folk interested
in attending. Let me know pronto if you’d like to know more.
PowerQuery is also great at
scraping data from semi-
structured websites like this…
…and serving it up into
interactive reports, like this:
PowerQuery/M and PowerPivot/DAX work hand in hand.
…clicking Refresh
Power BI dashboard example (Work in progress)
…divided by headcount.
Female only
Female only
All (Male + Female)
Highlighting on Maori Ethnicity vs dedicated Maori/Non Maori series
Be careful: It’s possible that people could misinterpret what highlighting shows,
as in this case where I’m highlighting on ‘Maori’ ethnicity.
Power BI offers lots of useful chart types. (And a fair few Excel offers lots of useful chart types. (And a fair few
useless ones too…) useless ones too…)
Power BI has awesome data handling (PowerQuery) and Excel has awesome data handling (PowerQuery) and
modelling (DAX) components. modelling (DAX) components.
Power BI can be used to create awesome dashboards. Excel can be used to create awesome dashboards.
Power BI is often used to create awful dashboards. Excel is often used to create awful dashboards.
A lot of helpful literature on what could make the A lot of helpful literature on what could make the
difference is sadly unfamiliar to the people building difference is sadly unfamiliar to the people building
these dashboards. these dashboards.
The more time you spend thinking about how users will make
comparisons, the less time they’ll have to spend comparing.