Developments, news, promising topics about data visualisation

The whole title:
Big Data, Big Dupe – A little book about a big bunch of nonsense

It was and is much written, talked about big data in recent years. For high hopes and vague opportunities even more money flows. Here comes this little book – published in February 2018 – by Stephen Few just right. Fluid to read, it admits with the errors that are so popular today in connection with large amounts of data:

  • Large amounts of data lead to more information only due to the quantity.
  • It counts the correlation, causality is negligible.
  • Statistical samples are obsolete due to the large amount of data.
  • We should measure everything, because we have data from everything. And a few more mistakes …

The exciting part of this text lies in the way these errors are revealed. Few refers his arguments to original statements by experts, quotes scientists and questions an entire industry based on concrete statements.


ConclusionCover: Big Data, Big Dupe - a little book of a big bunch of nonsense von Stephen Few
In short: Worth reading!!!

Slightly longer: It is not a clearly structured textbook, which one must work through. On the contrary, it entertains and provides interesting food for thought through many cross-references, links tips and hints. For those who are deep in the matter, it may be not detailed enough, leaving too many questions unanswered. But it helps to stay awake for the topic, it cleans up in a fun way with big data nonsense. The hope of Stephen Few is that if it´s telling the truth, something small can expose and unmask something great. In this sense, this small format serves the purpose.



For those Interested:

Title Big Data, Big Dupe
Subtitles A little book about a big bunch of nonsense
Author Stephen Few
Edition Published February 01, 2018 (available in English)
Pages 96
Publisher Analytics Press, New Jersey
Price about 10 $ depending on the provider
ISBN Print: 978-1-938377-10-5 (Hardcover); E-Book: 978-1-118-85841-7


If you want to read more about the author:


Management accounting analyzes, plans and informs. This requires a reliable database, but also flexibility in designing reports, presentations and dashboards. These requirements often can´t be combined in a software system. The discussion about the authorization of local reporting solutions, that complement central standard systems, is currently up-to-date. The article reports from practice and conveys the advantages and disadvantages of local solutions.

Central versus local reports

When I wrote ‘local report solutions’, I could already hear the prompt “think big or go home”. It seems that the management accounting is focusing on big data and the big BI solution. But what is the true situation? What experiences we make in our projects?

The utopia of the one system into which all the data flows and from which rapidly different reports can be generated to steer the entire enterprise, is the driving force of many positive, important developments. But the resource expenditure behind this idea ist enormous. This type of reporting is made practical by a high degree of standardization and continuity. Data collection and storage benefits from the general conditions.

However, the reports and presentations that provide centralized systems are limited by these conditions. They are not very variable and can only help partially in answering non-standardized questions.

Automated, local reporting tools can overcome these communication hurdles.

Therefore they are a flexible and cost-effective alternative. From our projects we know that a mix of central and local reporting solutions is the best course of action.

The local solution isn´t so local

At this point it is important to split the management accounting IT. The topics of data collection and storage are to be separated from the communication of the results, the reporting. Collecting data in isolated applications must be called into question for many reasons. Each BI vendor can name these reasons. However, automated local tools solve many issues when creating reports and presentations.

Strictly speaking, local reporting tools are hybrids. The data they access is usually from centralized systems. The reporting is done in front-end tools that are individually designed and managed locally. The distribution of reports can also be automated and is not particularly limited.

Sind lokale Reporting-Lösungen sinnvoll? Blogbeitrag auf

The right time for individual reporting tools

Use custom reporting tools when management often needs to change report variants. They make sense if structured data from the DataWarehouse are to be supplemented with unstructured, individual comments and explanations.

In strategic management, variable reporting and presentation tools make sense in every case. But operational management also raises questions that can only partially be answered by centralized reports.

Resource requirements

There is a catch. As a rule, locally generated reports can´t created quickly. The efficient creation of local reporting tools usually requires a very long learning period.

The design and planning process requires fewer reconciliations and compromises between the report creator and the report recipient. That’s positive. The technical and creative implementation requires a lot of know-how. Specialized knowledge is needed that has little to do with management accounting in a business sense.

Knowledge of the data connection, automation and programming are essential. If the visual presentation and presentation should be efficient and at high quality standards, an above-average user knowledge is necessary.

An idea, if own resources are missing

But even without a resource pool of its own, the advantages of local reporting tools are easy to exploit. The development of such tools is an individual service for which there are different solution providers on the market. These are not software providers, but experienced service providers with special management accounting know-how and relevant experience in the field of reporting.

Good service providers provide a very high quality standard that meets your individual conceptual and technical requirements.

They work directly with the specialist departments in manageable project periods and deliver immediately ready-to-use solutions.

chartisan is one of these solution providers. Our focus is the Management Information Design with IBCS® and the technical implementation with the possibilities of Microsoft Office (Excel, PowerPivot, Power Query, PowerPoint) and Power BI.

Advantages and disadvantages at a glance

You have little time? Therefore is here a clear plus-minus list:

+ Local reporting tools allow variable reporting and presentations.

+ These reporting tools can efficiently display structured and unstructured data.

+ Local reporting tools are ready for use in short time.

+ Through the high degree of automation, the tools support the operative controlling routine.

+ You can buy suitable solutions externally as a targeted supplement to the standard systems.


– Through internal development you tie up important resources, time and financially.

– When using external service providers, you buy foreign competences.

– External service also costs money.


Are there any questions left? Need a hint? You are welcome to leave a comment here or to contact me confidentially by e-mail.

Happy reporting,
Yours Silja Wolff


Data visualization in general means to bring abstract data and relationships into a visually comprehensible form. So far so good. But what are the unfamiliar forms of representation? When do I apply this? When are other forms more meaningful? I would like to introduce the project of Severino Ribecca to all those who ask these questions more often:


This online catalog is a library of various information visualization types. Initially, the project Ribecca served to expand its own knowledge through data visualization and as a tool for one’s own work.

He himself writes about the project: “However, I would like to know how it is. Although there is no such thing as a visualization method, it is not the only way to make sense.

And, fortunately, Severino Ribecca does not hide the knowledge, right on the home page is in the middle:


In addition to the first overview, the “Search by Function” option is particularly appealing to practitioners. Here again, the initiator relies on a fast-to-capture image language:

As soon as you click on an icon, you get a well researched and edited knowledge about the selected visualization type. Each display form is displayed with its functionality and application possibilities and is visualized with dummy data. At the lower end of the description, the reader also finds references to similar display forms:

So there are a maximum of three well-structured clicks up to the knowledge about individual visualization forms. I find this project visually very successful and meaningful. Perhaps this tip gives you some inspiration for your information visualization?


In any case, I wish you Happy Reporting 🙂

Yours, Silja Wolff





The origin of portfolios is in finance and describes a planning method of compiling a package of security papers (security holdings) dependent on to the criteria of return and risk. Later, this method has been applied to other areas in the 70ies. Since that time the portfolio analysis has been modified in many cases and is one of today´s widespread analysis and planning instruments of strategic management.  The main idea of each portfolio is the segmentation and evaluation of data volume.


Basics of illustration

Most portfolios are illustrated two- or tree-dimensional. With three-dimensional illustrations the point size  is the third value perspective next to the axes. You hardly see four-dimensional portfolios. Here the bubbles are segmented as circular chart next to their value based figure.

Basic requirements to all illustrations is, as always, the efficient legibility. This is manageable due to a fast visual comprehension to evaluate data by colouring of bubbles or using diverse symbols. Also marked segmentation criteria, such as lines, will support comprehension.


Scaling and value intervals


Other than the typical bar and variation diagrams the scales do not have to start necessarily with zero.

A deliberate choice of value interval is necessary, if the segmentation should be demonstrated reasonably.

In this context reasonable means all points are visible and do not waste space/room.

The more frequently a portfolio with diverse data sets is used, the more important is the deliberate choice of the scale. Note: often times a change of the software automation is necessary, as automated portfolio axes always start with zero in Excel. This is less beneficial if all of your data is between the value of 350 and 480.



Ledgends and captions

A special attention should be placed on the topic legends and captions. Mark/write directly at the relevant diagram components. That will increase the legibility immediately. Should you make the decision of using a three-dimensional portfolio, the value has to be written directly at the bubble as there is no scale for it.

Strategic management is unimaginable without portfolios. Every evaluation, every portfolio approach consists of strengths and weaknesses. But irrespective of the type and structure of each portfolio, the visualisation is key for a beneficial usage of this analysis tool. The mentioned universal criteria above are a good frame for portfolio illustrations.



Im letzten Blogartikel habe ich es angekündigt. Hier also mein Eindruck der druckfrischen Ausgabe von Stephen Fews neuem Buch:

„Bessere Entscheidungen können nur durch besseres Verständnis kommen. Informationstechnologien sind nicht der Schlüssel zu besseren Entscheidungen, das sind wir. Technologien können unser Denken erweitern, aber nicht ersetzen.
Computer haben ein neues „Datenzeitalter“ eingeläutet. Das „Informationszeitalter“ wird erst noch kommen. Wichtiger als das Sammeln von Daten ist es, das wir lernen Signale vom Rauschen zu unterscheiden.“ (Auszug) So fasst der Autor selbst den Inhalt des Buches auf der Rückseite zusammen.

Nach seinem letzten Buch „Information Dashboard Design“ beschäftigt sich Stephen Few jetzt mit der Grundlage guter Visualisierung, den Inhalten. Wie kommen wir zu guten Inhalten? Wie erkennen wir aus einer Fülle von Daten jene, die Entscheidungsrelevant sind?

Warum wäre eine „Slow data-Bewegung“ ganz nützlich?

Stephen Few teilt diesen Weg im Buch in zwei Hauptteile. Im ersten Teil geht es um Grundlagenwissen über Daten und Korrelationen. In einzelnen Kapiteln verbindet er bekannte statistische Aussagen mit betriebswirtschaftlichem Denken: Korrelationen von Kategorien, Maßnahmen; zeitliche Veränderungen; Zusammenhänge zwischen mehreren Variablen und Perspektiven. Die Darstellungen und ergänzenden Beispiele sind gewohnt einprägsam.

Als Begründung für das ausführliche Grundlagenkapitel nutzt Few ein Zitat von A. Lincoln: „Give me six hours to chop down a tree and I will spend the first four sharpening the axe.” Und fügt an:

„This is unpopular but sage advice.”

Der zweite, kürzere Teil des Buches liefert Ideen zum Installieren von Signalgebern, beschreibt wie durch Dokumentation eine nachhaltige Filterung des „Datenrausches“ nach Signale aussehen kann. Ergänzend hierzu finden sich am Ende des Buches zwei Praxisbeispiele.


Edward Tufte fasste bereits 2006 die Erfahrung vieler in den Satz: „Design cannot rescue failed content.“ Das vorliegende Buch von Stephen Few beschäftigt sich auf gelungene Weise mit den Wurzeln guter Visualisierung.

Auf unterhaltsame, nicht aufdringliche Art vermittelt Stephen Few mit anschaulichen Beispielen eigentlich trockenes Statistikwissen. Inhaltlich sollte denjenigen, die sich täglich mit Datenanalyse beschäftigen, alles bekannt sein. Doch die gezogenen Vergleiche, Darstellungen und Fragen sind gut geeignet, das unternehmerische Denken zu schulen.

Nicht für jede statistische Korrelation besteht auch ein handlungsrelevanter, kausaler Zusammenhang. Um dies nicht zu vergessen, empfehle ich dieses Buch.

Für Interessierte

Titel Signal
Untertitel Understanding What Matters in a World of Noise
Herausgeber Stephen Few
Auflage erschienen Juni 2015
Seiten 209
Verlag Analytics Press,
Preis 43,15 €
ISBN 978-1-938377-05-1



What are Bullet Graphs and where are they used?

Key data overviews on dashboards are easier to understand by visualisation. The realisation is quite difficult, because standardised diagrams are uneligible for this function. Bullet Graphs, a special type of bar diagrams, have been especially developed by Stephen Few. Basically they are very suitable for  those purposes, but they hold some disadvantages in its common form (see illustration 1 and 2), which we have solved in our projects due to a modified, optimised form of visualisation.


Illustration 1: The components of a Bullet Graph as it is regularly used in dashboards

  • Bullet Graphs are a special form of bar diagrams
  • They compare a presented key figure with a target value
  • By using background colours, additonal information of quality will be allocated
  • Bullet Graphs are often used in dashboards. They save space and replace tachometer
  • Bullet Graphs can be aligned horizontally or vertically


Illustration2: Dashboards with Bullet Graphs as they are often recommended and used


 5 disadvantages of this Bullet Graph illustration

  • The graphics become confusing very quickly, because many detailed information is shared, which is not necessarily needed or partly not available at all. For example the quality statement “bad, okay, good”. This disadvantage mainly occurs when many Bullet Graphs are displayed on one page (list of key data).
  • Variations of the goal are not visualised and therefore difficult to see. The only part easy to read is, if the bar of the just value is overstepping the target value. For the viewer it still would be important if the overstepping is good or bad and how large the variation is.
  • Coloured background areas distract the viewer from the main information, which is the bar of the just value and the target value. The coloured areas always look the same, no matter if the goal was reached or not. They rather disrupt the perception than to support it.
  • Background areas decrease the contrast towards the bar of the just value and target value. Thus, these essential graphical elements are difficult to see in the front.
  • Target areas are difficult to illustrate. In practice there are not only set target values (single values) for key data but often also for target areas. With this key data (KPI) the goal is considered achieved as long as the just value is between the minimum and maximum level of the target area.


How to ideally use the idea of Bullet Graphs in dashboards

For the use in dashboards we use a specially optimised form of Bullet Graphs (Illustration 3), which avoid the named disadvantages and bring additional visual features.


Illustration 3: Dashboard with Bullet Graphs recommended by chartisan


  • The simplified illustration without unnecessary background areas allows a quick overview
  • Indicated are the target values (single values) or target value areas (minimum/maximum level)
  • Variations will be highlighted in colours: positive (desired)= green, negative (undesired) = red. The extend of the variation is easily captured. Viewers immediately see where there is “a lot” red or green flashing.
  • The value caption is space-saving and clearly stored within the columns of the chart.
  • This optimised graphic form uses the strengths of the original Bullet Graphs and complements them with a visual drift indicator. The detail level has been reduced as much as possible, whereby these optimised Bullet Graphs are ideal for lists of key data.


By the way, we have managed and automised the displayed example with some tricks in Excel. As often the idea was key not the tool. Having said this, we hopefully could inspire you. Should you look for further good ideas for your reporting, contact us!


Ein kleiner Projekt-Einblick in ein übersichtlich aufgebautes Sales Dashboard. Durch die Realisierung in Excel, ohne etwaige Plugins oder VBA-Programmierung, lässt es sich komplikationslos von jedem Office-Anwender nutzen und verteilen. Die Datenaufbereitung basiert in diesem Beispiel auf Pivot-Tabellen. Eine Datenbank-Anbindung ist problemlos möglich. Das ist keine Zauberei oder ewige Tüftelei. Versprochen! Wenn Sie unsere Lösung interessiert, […]