13.4.12

Welcome to Better BI's new home.

I've recently moved the Better BI blog here from its old home.

The reason: I can embed Tableau Public-published Tableau content directly into this blog. The old blog didn't have this ability and the workarounds weren't acceptable.

10.4.12

Finding Poetry in Data - Revealing the Rhythm of the Tides with Tableau

We don't normally think about BI being the path to beauty, but sometimes the practical pursuit of information takes us to unexpected places.


Tide Time

The tides do not tick-tock
to our diurnal clock
in their steady march
across the times of day.

Our daily rhythms are joined
in rolling harmonies;
today the rocky spit
is covered at morning coffee,
tomorrow the waves break and foam.

The ocean's song evolves,
the earth, moon, and sun
tuning the great liquid bell
in geologic time.

Two months on the Panamanian Pacific coast. The beach is right below us, the surf a constant presence.

Living right on the ocean has it’s great benefits. The beaches are strikingly beautiful; the surf intermingles black volcanic and white sand into fantastic patterns visible from space (or at least in Google Earth and maps). The ocean has shaped the beaches into a steeply sloped upper section, and a nearly flat lower zone wonderful for all types of activities. A time lapse video of the receding high tide in the early morning is above.

It's hot. Not absolutely blisteringly hot, but hot enough to avoid strenuous outdoor activities during the middle of the day.

The tides are quite high, averaging almost 15 feet. It's important to know the tides so that you can plan your beach activities.

Fortunately, the local community has posted the tides online in a series of web pages, one for each month; March, 2012's page is here.

Unfortunately, the tides are presented in a table that, although it contains the data, is minimally useful. As shown for April 2012 (partial) below, in addition to the obvious analytical impediment of it being a table, each Months' tides are presented on a separate page, and the time and height values are commingled in the table's cells.


I was faced with the classic BI situation: plenty of data, little insight. My challenge was to come to grips with the data and render it in an engaging, approachable form that provides the information people need.

What to do? What to do?

The usual: a little creative data munging to pull the tides data out of the HTML tables (straightforward stuff) and then a little exploratory data analysis with Tableau to see what the data tells us, and how we can create some visualizations that communicate the information we (and others) need.
The data munging stuff is interesting. There are a number of ways available for pulling data out of web pages for Tableau analysis. In straightforward cases a simple copy-and-paste works wonderfully well. Beyond that it can get very interestingly complex. But that's for another post, this one's about poetry and beauty.

Rhythm and Pulse of the Tides

I grew up in Shelburne, Nova Scotia, with one of the best harbours in the world at the end of my yard. Next door was a shipyard where the last of the wooden open ocean commercial fishing boats were built by men using centuries-old traditional methods.

My boyhood was spent at the ocean's edge. I was a pirate, sailor, fisherman, an explorer sailing off for exotic lands.

The tides were as much a part of life as mealtimes, school hours, and all the other regularly scheduled daily activities. Knowing when the tides were high and low, and when to expect them to be bigger and smaller mattered. For me it controlled the size of my playground, for many people knowing the tides was central to their existence.

People born to the ocean know the tides. They don't need graphs and charts and tables and schedules. The tides infuse life, their rhythms organically intertwined.

Here on the Panamanian Pacific coast the information about the tides is much for more prosaic uses, at least for the recreational ocean users. At low tide the beach is available for all manner of recreational activities. At high tide there's much less opportunity, e.g. it's really difficult to run on steeply sloped soft sand. At very high tides there's virtually no beach.
Knowing whether you'll be able to run on the beach at sunup, after coffee and before breakfast, is nice.
Knowing where the tides will be for this week, next week, next month, helps to plan your activities.

How to help people understand the tides? Understanding them not as a list of figures in a table, but as the dynamic, regular, but variable system they are.

How to help people internalize the nature of the tides, so that they can adapt to their ebbs and flows, risings and fallings? Their rhythms.

Enter Tableau

Once the data's available it's straightforward to do some basic visualization with Tableau. Straight away it becomes obvious that there are regular patterns to the tides. High and low tide times move forward from day to day, the time from high to low a little over six hours.

What We See

The tides are the long, slow, resonant beat of the gravitational dance between the Earth and the Moon, with harmony from the Sun. The dance is eons old; the partners know each other well, their movements intimate and comfortable.

Visualizing the tides with Tableau clearly reveals their dynamic structure. Although there is a regularity to their oscillations they are not strictly mated to the cycles of day and night. The tidal patterns stand out clearly in these visualizations, confirming and exposing our innate sense of the tides' haughty disregard for our clocks and convenience.

Daily Tides

Daily Tides shows the tides wrapping themselves around the days in a double helix. The large scale structure mirrors the Earth's rotation, the tides the ocean's response to the moon's beckoning.

The vizualization clearly maps the tides to the 24 hour day, advancing by almost an hour from each day to the next. If low tide at sunup is your pleasure, you're in luck—it comes around every sixth day. If you like high tides, they occur a little over 6 hours apart.

Tide Height

In their celestial dance the moon swings close, pulling the tides higher, then steps back, letting the tides subside. All the while the Sun smiles, attracting and influencing the oceans—sometimes in step with the Moon, sometimes not.

The polyrhythms of the dance—Earth's rotation withn the Moon's orbit, both in Solar orbit, show up clearly in the eternal cycling of the tides high then low then high.


Earth - Moon - Sun Tidal Relationship


source: Wikimedia

4.4.12

BI is not an IT activity.

The gulf between the information needs of real human people and IT is almost universally experienced. It's helpful to recognize the drivers for this situation in order to attempt to correct it. Some of these have been mentioned in previous posts.

There's a fundamental, unrecognized, nearly fatal flaw in the ointment, revealed in the term "IT" itself: the casting of information acquisition, management, understanding, use, and communication as a technological activity, hence "Information Technology".

It's not.

Information is a human property. Casting it into the realm of technology presupposes that it's a physical resource that can be effectively processed with an industrial assembly line approach, given the acquisition of the appropriate machinery, the installation of an adequate industrial-production control model of oversight and management, and staffing up with the appropriate level of resources (not people, revealingly). Once everything's set up, the big green "Go" button can be pressed (metaphorically) and he wheels will turn, the gears will mesh, the data will be pumped in, and the information the business decision makers need will come out.

If you've been underserved by your IT department in getting your questions answered, your information needs satisfied, almost everything you've experienced, all the trouble getting results from IT, is caused by the application of this paradigm. And almost all the solutions suggested are applications of more of the same to address the weakness inherent in the model.

One example: the creation of a Business Intelligence Center of Competency (BICC) as the solution to the problem. It can work, but only if it's not just the addition of another mechanical/operational command-and-control structure with the intention of assuaging the pain causing by the existing environment.

In reality, too many BICC initiatives start out well, then crumble into the same industrial dust as the IT departments whose functions they're replacing. As soon as friction sets in, and it will, usually in the form of someone important not getting something they want as soon as they want it, the impulse is to add mode control systems to manage the submission of requests, the analysis required to understand them, and the prioritization and queuing systems to manage their satisfaction.

All of this misses the essential, central truth: business intelligence - the analysis of business data in order to achieve information and insights from it, and to understand the stories it has to tell, are human cognitive and intellectual skills.

It's important enough to repeat: BI requires human thinking skills. It's not an industrial process.

Business people who need to understand their own data understand this, which is why they use Excel, Access, and increasingly, the new (now 5+ years old) generation of direct-access, low-friction highly effective data analysis and visualization tools to help them understand their data. And they're increasingly frustrated with their IT organizations, which don't get it.

Recasting the concept of BI into its proper form is an essential element in solving this problem. Layering on more industrial management processes only compounds it.

Successful BI occurs when it's recognized as a professional practice, requiring of its practitioners the appropriate combination of aptitude, interest, knowledge, training, and experience. Then the barriers can come down and the myriad distances between real human people and the business intelligence locked up in their data can be freed.