Share your ideas for the JMP Scripting Unsession at Discovery Summit by September 17th. We hope to see you there!
Choose Language Hide Translation Bar
How to make practical sense of data (and win a book)

Kaiser Fung is a statistician with more than a decade of experience in applying statistical methods to unlocking the relationship between marketing and customer behaviors. He leads a team of statisticians at Sirius XM Radio responsible for gaining insight into customers and operational best practices.

You may know him from his popular blog, Junk Charts, which pioneered the critical examination of data and graphics in the mass media. He was also a keynote speaker at the JMP Discovery Summit in 2010, at which he gave a speech that asked “What Happens After the Math Is Done?” And he is author of the book Numbers Rule Your World. (We've got 25 signed copies for a special JMP Blog giveaway. Details are below.)

Fung believes the analytics community should spend more time thinking about how to most effectively support data-driven decision making -- and he explains this point in a new story "On a mission to make more practical sense of data":

    Information-design guru Edward Tufte threatens that every time someone gives a PowerPoint presentation, he will kill a kitten. Kaiser Fung knows Tufte is only kidding, but Fung is still concerned. 

Fung has seen too much not to be concerned – too many slides jammed with absurd amalgamations of information.

There’s so much information to ponder, so readily at hand. But insights are often lost in the clutter or lack practical value.

Don’t get Fung wrong: He is in favor of data and data-driven decision making. He worries about how presenters lose the audience when conveying quantitative information.

“A lot of things are needed in order to bring insights about numbers to the point where you can get nontechnical people to move in a direction that is essentially driven by evidence and not simply by data,” Fung says.

Read the whole story, including a sidebar about the role of JMP in Fung's work.

And if you are among the first 25 people to comment on this blog post explaining your strategies for promoting data-driven business decision making in your organization, you could win a signed hardcover copy of Fung's book, Numbers Rule Your World. Your contribution to this discussion should be between 50 and 75 words long. Be sure to enter your e-mail address when you write your comment so we can contact you if you are a winner. Only one book per commenter.


Sumanth wrote:

I have tried to push myself and my co-workers to question the decision making process if it is not backed by supporting data and statistical tests


Carolyn Nielson wrote:

I think of graphing as visual storytelling. It is similar to good flash fiction where a story is told concisely and yet tells a complete story with emotional impact.

With a graph, the main message should be clear to the viewer at once, perhaps provoking questions which require a more careful study. A longer perusal reveals more information, if properly rendered.


Damon Edmondson wrote:

Regardless of winning or not...

its often a matter of showing hard data points. But I have found its first important to find key leadership staff that believe in the point the data supports then to present the data to all.

Data on its own can often be disbelieved, regardless of their truth. Aldous Huxley said, Facts do not cease to exist simply because they are ignored.


Dwinanto Santoso wrote:

Few things that I think believe can help my as an analytic in order to create value & support the management and would like to share with the community:

1. Get you internal client engagement as early as possible, especially in the data collection.

2. Understand what the decisions that they have to make are, decision criteria & constraint. We need to follow their guideline as at the end of the day the need to believe that it was their decision.

3. never start with a number, number make people alert and might be defensive. If people want to get the hard fact they can get it in spreadsheet later. make sure that we plotted the story right.


Chris McGraw wrote:

I work for a data service provider for an emerging form of commercial real estate, where good data and analysis is few and far between. Since we are the only providers of reliable data it is our responsibility to provide an unbiased snapshot of fundamentals in the market place. I've found that to do this for aggregated data you need to present more than just what the top line figure is doing. Commonly we provide data on the distributions of data - noting outliers, and color on how the numerator/denominators are individually behaving if we are working with a ratio series. We want to give the full picture and be able to why metrics are moving in the direction they are.


Gary T. wrote:

Ask the question that you will answer with data.

"Do we get enough traffic to justify the expense of continuing a web application?"

"Did we make money from a Groupon offer?"

Collect the data, perform the analysis, and produce graphs. Although numbers are useful, executives respond to pictures.

Ultimately, one needs to show value from data analysis. More often than not, that value needs to be financial.


Meic Goodyear wrote:

In public health there's a mass of data, not always what we want, and not always complete or of the highest quality. I try to make the presentations as comprehensible as possible for a given audience while still being statistically sound, to indicate explicitly the limits of uncertainty, to incorporate geographical presentation wherever possible,and to seek out meaningful patterns collaboratively with front-line experts.


Daan wrote:

The railway maintenance industry is going through a transformation, led by a deluge of data. With new measurement tools, more (up-to-date) information on the railways, switches, power lines, &c. is becoming available. Combined with the right analysis, this is now being used to fundamentally change the approach to defining not only maintenance schedules, but also restructuring maintenance contracts, KPI's, research demands, and more.


Antonio Rinaldi wrote:

In my opinion the key point is: simplification, that is, reduction of the problem to basic blocks. A statistician can develop any rather complex system, but he also has to make himself understoodwhen and make the point clear to other people.

Any effort to simplify the problem should help the process of carrying out the most convenient type of data analysis and then take the best decision.

P.S.: please discard my previous post, I have made a typo error in my email address.


Lisa K wrote:

I am the web manager for a secondary college. For an organisation who measures success by numbers, you'd think that getting Dept Heads to make decisions on web strategy and content, based on data collected from website visitors would be easy. I've found that this approach works well:

1. Explain the source and reliability of the data

2. Present the data in overview format first, then drilling into detail to show trends over time, and specific interest in certain sections.

3. Present relationships between different events on our calendar, and spikes of interest in our website.

4. Support this with anecdotal evidence or data from interviews.

5. Use this to focus attention on areas that are performing well, or need more TLC.


Andrew Marritt wrote:

My audience are HR professionals, a group often not highly numbers-driven. The starting point always has to be the decisions that they need to make. This, and only this needs needs to guide how information is presented. Too often there is the temptation to present all the data, or interesting patterns in a way that numbers-literate can understand but this can dilute the key message or confuse. You need to resist this temptation.

For regular reports for larger audiences we typically use usability testing techniques to observe how people use them, where they look, how they interpret the information, what other questions are triggered.


Shawn Shafer wrote:

Consistency has been the key for me - being a ready companion, consultant to explore the data and extricate its meaning as well as predictable work products themselves that business leaders rely on and plan against. After establishing an evidence based decision making process in one key area, the case for expanding that process is largely done by the executives looking for it.


Andrew wrote:

The way I have gone about driving a culture of data driven decision making in the organisation I belong to is to:

- Show the benefits through quick wins

- Enable people to own the process of data collection to improve the quality of the data

- Encourage people to understand how their data fits into the overall big picture

- Enable people to critique reports/charts and ask honest questions


Edgar wrote:

To promote data-driven business decisions, enterprise need to hold more townhall style meetings. they need to disallow the use of laptops or social media while people are talking and pay attention at what is being said so that - as people share ideas, concerns and complains -- everybody is aware of those and creates strategies based on the information give. Our lack of data-drive decisions really stems from people don't knowing stuff about the business.


Glenn Waddell wrote:

I am a high school AP Stats teacher who uses JMP in the classroom. In addition, by asking for data from our student support offices, our testing data, and other data generated by our school and then analyzing it with JMP, I have been able to show our administration that some of the preconceived notions they had are wrong.

This takes my school from being data driven (through excel graphs, ick) to being data informed so we can make good decisions. Thank you JMP!


Keith Suckling wrote:

The major challenge for anyone presenting information and the insights from it (note: not analysing and reviewing information in a workshop setting - that's different), is the seeming contradictory yearning for simplicity combined with the tick the box mentality requiring the detail be visible.

Is it implicit distrust? Sometimes.

Maybe we as practitioners need to be stronger in maintaining different vehicles for different information. Any PowerPoint guru will have something like "One slide, One point" in their list of guidelines. Maybe we should hold ourselves to the same thing when presenting and save the clever analytical graphic that has heaps of messages for another medium.

"What you leave out is just as important as what you leave in"


Patrick B wrote:

One thing I try to do is to follow Einsteinâ s advice: â keep things as simple as possible, but not simplerâ . I find that this is especially important when trying to influence people because many decisions are made very quickly, possibly just on first impressions. If people can look at something and quickly get what they need from it, theyâ re more likely to be moved in the intended direction.


annonymous wrote:

First and foremost, I've been dragging everyone I can find to Edward Tufte's seminars for years now. Seldom does anyone come back without dropping at least a few bad habits. I also added the Junk Charts and Numbers Rule Your World RSS feeds to our intranet homepage.

JMP has been a great tool for us because it allows everyone to EXPLORE the data we gather, not just passively receive it. It's in the exploration that we all learn and develop our skills.


Ben Spigel wrote:

One of the most useful applications of data visualizations is to make clear ongoing mistakes that just seem to make sense. The biggest issue is selling something (either a product or service) as a loss with the hopes of getting new or repeat customers. This has become even more critical with the advent of Groupon, which makes a pitch that makes sense on a visceral level (deep discounts will bring in lots of people), but require deep tracking and number crunching to show if it will actually be profitable. Visualization is key to understanding why something that makes sense today might not work tomorrow.


Eric wrote:

Trained as a physicist, I am accustomed to letting data drive decisions. Iâ m not a lawyer, and have found that in the professional services industry (or maybe just the legal field), data is far from king. This is true both in the management of a law firm, and also when presenting data about a case (to a jury, etc.). One of my goals is to fix this by applying Tufteâ s principles to make the relevant, interesting, and important data jump off the page (or screen).


Mark Neely wrote:

My strategy to create a culture of data-driven decision making is to team with others and explore the data dynamically looking for patterns, trends and correlations while trying to make the most out of the data at hand and through active experimentation (DOE or BACI designs). Involving others while analyzing the data using the many visual / exploratory platforms available in JMP allows others to discover relationships that would have been completely missed otherwise.

When they are part of the analysis and are there when something important is discovered, then they are more interested in actively participating in the future (as opposed to them just being given the results).


Sushil B wrote:

Pointing out that the effect of varying factors in a DOE is about the same as the natural variation in the process has prevented some costly process changes from being made.

In my view, these kinds of engineering decision are business decisions since they directly affect the bottom line.



Kun Wang wrote:

As an independent pharmaceutical freeze drying professional consulting organization, we deliver our customer best practice in freeze drying R&D, production operation, labor force management and equipment maintenance. JMP help us daily to collect data, perform root cause analysis, discover opportunity to help our customer do more with less. Our JMP based GMP Visualization System help our customer use the least resource to improve pharmaceutical freeze drying product quality to meet compliance requirement. Also data driven based GVS help customer remove operation bottleneck to obtain productivity increasing.


Sizhen Wang wrote:

During my 6 years career as process integration engineer in semiconductor industry, I also find the presentation of experiment data analysis is much more important than the data collect, digest and analysis steps. more fresh or junior engineer try their best to convey the detail information how they conduct experiment and do analysis, and throw much junk words in presentation, the disappointed things are in most cases, boss or supervisors don't buy-off, and request further do or re-do the job, which make engineers frustrated.

while the senior technique expertise conduct the information in a total different way, they just tell a story with simple word as much as possible, they make decision and try to persuade management which option is reasonable and practical, not let boss to select as new engineer usually do. and this way is also very effective, it reduce the decision risk and workload of management as it is you which is familiar with the details.


Andrew Landgraf wrote:

One of my main strategies is to display the data in a way that is easy to interact with it. Whenever I present, there is always a question that comes up that I didn't think of. Making the data/chart interactive, means I can easily answer many more questions than static graphs. Also, if it is simple and intuitive, I can let the decision maker change the parameters to test any theories he or she has.


Steve Reagan wrote:

Of course, building DD decision making into our [actual, relaistic] business practices is an ongoing challenge. There extist many, encumbent mental heuristics that are not easliy dislodged. Of course, this is human nature.

The three areas I HAVE been successful in using data to change views are:

1) Newbies, even management ones. They don't have any inter relational hangups to overcome and are MUCH quicker to reach true convictions. The issue here is they are hard to come by and don't last long, so grab 'em while you can.

2) Persistance: Develop a data based report, presentation, etc. with a clear and simply translated message and stick with it. Update the dta aas it becomes available, but stick to the theme.

3) Be truthful, logical, and sincere. It's your data, no one knows it better. It should be logical to be followed by non-specialists, and if you really believe in it, others will know.

My thoughts for today ...



Christopher K wrote:

I'm working as a black belt supporting the high impact projects and find the use of statistics is something to be improved upon. My strategy is to utilize the visual powers of JMP to explain the data in the form of a story and provide simplified figures that point to clear direction so the business leaders and lead scientists/engineers can focus on creating the solution rather than spinning their wheels with the assumptions. A lot of my work is focused on creating structured DOEs and Visual Interpretation.


Danny Kugler wrote:

At Hewlett Packard our presentations are expected to be in Power Point so I canâ t do much about Dr. Tufteâ s threat. We are however encouraged to constructively critique our colleagues. The author and audience be they management, marketing, or engineers are usually from engineering backgrounds so they are keen to make data driven decisions. Graphics that get in the way I can help correct via my usual practice of "helping" the author re-present the data without the usual chart junk.


Arati Bechtel wrote:

Hi, Sushil,

Thanks for your comment. We are particularly focused on business decision making here -- how can the analytics community influence business decisions? That's what the story about Kaiser Fung is about. Do you have any examples like that -- where your engineering group was able to influence a business person to make a decision?


Sushil B wrote:

Working in an engineering group, it is easy to get people to understand the importance of using data to drive decisions. The difficult part is to get them to use the data correctly. What I find is that people for the most part don't understand variability. I am constant telling people the importance of randomizing runs in during a design of experiments, the need to run a center point multiple times to determine natural variance etc.


Lisa Haneberg wrote:

I think that one of the most important aspects of analysis is building trust. I want to establish myself as a reliable interpreter of information. And when I establish credibility, I can show less to convey more.

Every now and then we will get lost in the weeds, but more often, I am able to help leaders focus on the most salient findings because they know and feel comfortable with my process. I recall one presentation where I stated that I was not sharing XYZ because that was not where the story was and would be a waste of their precious time. These leaders expect this of me. They also know that I am willing to show them the nitty gritty any time.