Pattern-based visual decision making

I think that Edison said: innovation is 1% inspiration and 99% perspiration.

As over the next two years it is currently planned that I will not have time to do my usual consulting activities, beside for some activity online, mainly pro bono, I started over the last few weeks to share some material that I had planned to use in my services or with my partners.

Until recently, I shared ideas and concepts or blueprints with my network (albeit less and less frequently from 2006, when I started to prepare to resettle in Brussels), but I think that some ideas could be used for other activities before end 2011, and therefore here we are.

This short preamble will be in each article.

If you manage to turn something to a practical use- good luck: you will have to do the 99% missing!

And now, we can move to today’s article.

Pattern-based visual decision making

This article derives from some material that I already published online, first in 2003, and more recently here (the “Future of IT” series), and it is the first of a series of a little bit more practical articles on various subjects.

In both cases, the key issue is knowledge management and communication.

My concept of a pattern is, as I wrote before, a little bit wider than usual.

In organizations, I consider a pattern as a composite identity, whose components are both formal (e.g. the organizational structure) and informal.

Did you ever read something about vision in robots, like those machines that read the hand-written addresses on your envelopes?

When you search objects, or within a list, you do not read every letter- you skim through the text, and when something seems close, you look a little bit more.

In the late 1990s I created an application to manage the “infoglut”, i.e. to allow getting informed on what you need when you need it- and be confident that it is always up-to-date.

Eventually, I reused the logic to store knowledge about the structure of an organization, roles, offices, and so on, to support few customers.

A visual database

But storing the information about an organization, its members, their roles and tasks is relatively simple: it is a matter of storing and retrieving information that is structured.

Storing patterns in my case includes keeping information that is “fluid”- the organization as a set of interactions.

Few years ago, when Flickr, YouTube, etc became an everyday tool, also mainstream (i.e. non-technical magazines) started talking about a new issue: searching inside pictures.

What is usually available, is just a source by parameters- size, shape, etc.

But there is a small catch: you have to add that information along the picture or movie.

The interesting part is that technology is already available, if you want to do something more.

Nowadays, it is quite difficult to buy a mobile phone that does not come with a camera.

And cameras have often net autofocus systems that differentiate between people and objects.

Do not store verbal information.

Instead, convert verbal information into a visual code, and then use the existing technologies to search visually- as a person or using bits and pieces of technology.

Pattern-based visual data management

Do you know how the camera chip within your mobile phone works?

To make life simpler, let’s just say that each image is decomposed into components, and that electrical states represent the presence or absence of information.

Therefore, you could convert a “visual map” into what, in the end is just a single quantity.

But, of course, the key issue is: defining a “coding” to represent the knowledge into a visual map, that then is converted into a code, and then stored.

Retrieving information is simply a matter of nit-picking from a “dashboard” the characteristics of the “pattern” that you would like to build (organization, rules, etc), and then have the system search it for you.

The beauty? If you work with databases, you know that you have to enter one or more keys, and that the larger the number of keys, the longer it takes to find the information.

But if the information is stored to solve a specific “pattern”, it is just a matter of retrieving a specific code.

Much faster than the typical retrieval system that you see, say, in CSI to match a fingerprint, or DNA code.

And coding is the key.

Moving to visual decision making

Getting back to the beginning: we do not read a list item-by-item: we search to find matching patterns- what exist, and what we search.

If your coding system is properly designed, you can imagine different “visual maps”, and then ask the system that I described in the previous sections to convert these maps into unique figures, and retrieve the closest ones.

In the end, it is what you do with your dashboards, or all the DSS/EIS/Business intelligence systems etc that I worked on or designed from late 1980s.

The difference is: until now, you converted your patterns into numbers, retrieved numbers, created new patterns on paper, on a whiteboard, in your mind, and then, again, you converted into numbers, to refine your decision.

The funny part: you do the job to make yourself understood by the computer- and not the other way around.

I plan eventually to build and publish a prototype, but I am quite confident that I am not the only one who toyed in the 1980s with Artificial Intelligence, in the 1990s with decision-making technologies, and over this decade with various bits and pieces that try to simplify your interactions with technology.


The concept has few interesting consequences: the most important being that, while all the database technology is still there, you can now build a “layer” to make it more humane.

Instead of having to wait to have a pre-defined dashboard, you can visually explore, store, search information.

There are some companies offering software tools that, with some data-massaging and integration, allow to explore data almost visually.

The difference here is to build not a general-purpose, one-size-fits-all system, but to build, search, retrieve by patterns.

And patterns can also be aggregated with other patterns.

Nothing forbids the actual decision maker to “save” patterns that (s)he could then reuse, or share.

What is needed, is: work not only on the technological side (relatively easy), but also on the education of the users.

Current offers like GoogleWave or Facebook are mainly verbal- no surprise, considering that the founders received book-based education.

It will be interesting to see how YouTube and others will introduce visual technologies not only to retrieve information, but also to store it.

XXI century managers who started using mobiles over the last few years, and are now just kids, will probably better equipped to interact with a visual approach.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s