Visualization: taxonomy and dependencies

27 Jul 2007 - 8:59am
7 years ago
6 replies
730 reads
White, Jeff
2007

I'm working on possible visualizations for a large taxonomy my company
uses to describe our products. To give a quick example, say we sell
watches. There are lots of possible "facets" like brand, size, material,
etc. Each of the facets have values, so the brand facet would have
values like fossil, rolex, etc. Many people across the company can add
descriptive data about a product as it moves through our supply chain.
There is also a smaller set of users who manage this taxonomy. Those
people need to create, manage and view what we're calling
"dependencies". This is where my visualization challenge is.

A dependency in this case means "the selection of one facet value can
change the possible values of another facet". What my team is trying to
do is to come up with a visualization that will allow a user to interact
with a large data set easily and view the various relationships between
facets as well as established dependencies.

I'm having a very tough time finding any research or examples for a
situation like this. Can anyone help? I hope I've done a decent job
explaining the situation.

Jeff

Comments

27 Jul 2007 - 9:28am
leo.frishberg a...
2005

Unless I've mistaken you're need, take a look at Second Story's
work...I'm specifically thinking of their multi-d matrix for their
"collected works" - http://tinyurl.com/3d868x.

Leo
-----Original Message-----
From: discuss-bounces at lists.interactiondesigners.com
[mailto:discuss-bounces at lists.interactiondesigners.com] On Behalf Of
White, Jeff
Sent: Friday, July 27, 2007 7:59 AM
To: discuss at lists.interactiondesigners.com
Subject: [IxDA Discuss] Visualization: taxonomy and dependencies

A dependency in this case means "the selection of one facet value can
change the possible values of another facet". What my team is trying to
do is to come up with a visualization that will allow a user to interact
with a large data set easily and view the various relationships between
facets as well as established dependencies.

I'm having a very tough time finding any research or examples for a
situation like this. Can anyone help? I hope I've done a decent job
explaining the situation.

Jeff

27 Jul 2007 - 2:42pm
John Raisch
2006

What a great (and tough) question. We've titled that concept --
Predecessor Filtering.

We've also spent more time trying to address this in the ui for
consumers, and much less so for business users via our catalog
management system. (seems maybe like you are asking about this in
context of catalog management. To give some insight on that I'd need to
do some digging).

As far as consumer facing ui, here are a few visualizations.

Jewelry @ Zales.com. Browse to Shop By Product: Rings. Select Stone
Type: Diamonds. Notice Carat Range is now a facet (which was not
available prior)

Power Tools @ AceHardware.com - Browse Lawn & Garden > Outdoor Power
Equipment > Trimmers/Edgers/Blowers > Blowers. Notice there are 3 values
under Power System, underneath each is a set of distinct groups (gas has
engine size: 25cc, 31cc; electric has amperage:10, 12, etc...)

For consumers, we've kept it simple. I could see how visualizing this
for content mgt purposes would help communicate the relationships, but
could get ugly very quick...

Our next challenge is to apply Predecessor Filter concepts to our new
async browsing feature. If anyone has a visualization for that, pleas
send them along!

http://www.thesportsauthority.com/shoeFinder/

-----Original Message-----
From: discuss-bounces at lists.interactiondesigners.com
[mailto:discuss-bounces at lists.interactiondesigners.com] On Behalf Of
White, Jeff
Sent: Friday, July 27, 2007 10:59 AM
To: discuss at lists.interactiondesigners.com
Subject: [IxDA Discuss] Visualization: taxonomy and dependencies

I'm working on possible visualizations for a large taxonomy my company
uses to describe our products. To give a quick example, say we sell
watches. There are lots of possible "facets" like brand, size, material,
etc. Each of the facets have values, so the brand facet would have
values like fossil, rolex, etc. Many people across the company can add
descriptive data about a product as it moves through our supply chain.
There is also a smaller set of users who manage this taxonomy. Those
people need to create, manage and view what we're calling
"dependencies". This is where my visualization challenge is.

A dependency in this case means "the selection of one facet value can
change the possible values of another facet". What my team is trying to
do is to come up with a visualization that will allow a user to interact
with a large data set easily and view the various relationships between
facets as well as established dependencies.

I'm having a very tough time finding any research or examples for a
situation like this. Can anyone help? I hope I've done a decent job
explaining the situation.

Jeff

________________________________________________________________
Welcome to the Interaction Design Association (IxDA)!
To post to this list ....... discuss at ixda.org
List Guidelines ............ http://beta.ixda.org/guidelines
List Help .................. http://beta.ixda.org/help
Unsubscribe ................ http://beta.ixda.org/unsubscribe
Questions .................. list at ixda.org
Home ....................... http://beta.ixda.org

27 Jul 2007 - 2:58pm
White, Jeff
2007

Leo & John - thanks for the responses. As a display of my vast
humility & modesty: I did a very poor job of explaining my problem
:-)

The app we are (might be) building is not for external customer
consumption - it's not a way for a *PERSON* (ahem, thank you very
much) to navigate through a product catalog and find what they want.
It is a way for an employee to provide descriptive data about a
product, so it can move to various sales channels.

Someone manages the actual product taxonomy: the facets and values
under those facets. The dependecy comes into play like this:

We have two facets, buyer color and merch color. Buyer color is
industry standard and contains values like blue, brown, red, etc.
However, the clever folks in our marketing department don't really
care for those boring and very informative words, so they like words
like sky blue, ocean blue, etc etc etc.

So, if an employee has a product and is providing details about it:
they might select a buyer color of blue. This selection would limit
the values of the merch color facet to only those values associated
with blue. It wouldn't display leprechaun green, blood red, etc etc.

I hope this is making sense. The visualization challenge is how can
the taxonomy manager can navigate through this taxonomy and easily
view all of the relationships and dependencies shared among taxonomy
facets. The entire data set is quite large.

Sorry for the bad description earlier, I hope this one is better.

Jeff

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://beta.ixda.org/discuss?post=18738

27 Jul 2007 - 5:03pm
Kim McGalliard
2006

Would something like the Visual Thesaurus work? It's often what I think of
when I think of multi-faceted relationships.

http://www.visualthesaurus.com/

On 7/27/07, Jeff White <jeff.white at jtv.com> wrote:
>
> Leo & John - thanks for the responses. As a display of my vast
> humility & modesty: I did a very poor job of explaining my problem
> :-)
>
> The app we are (might be) building is not for external customer
> consumption - it's not a way for a *PERSON* (ahem, thank you very
> much) to navigate through a product catalog and find what they want.
> It is a way for an employee to provide descriptive data about a
> product, so it can move to various sales channels.
>
> Someone manages the actual product taxonomy: the facets and values
> under those facets. The dependecy comes into play like this:
>
> We have two facets, buyer color and merch color. Buyer color is
> industry standard and contains values like blue, brown, red, etc.
> However, the clever folks in our marketing department don't really
> care for those boring and very informative words, so they like words
> like sky blue, ocean blue, etc etc etc.
>
> So, if an employee has a product and is providing details about it:
> they might select a buyer color of blue. This selection would limit
> the values of the merch color facet to only those values associated
> with blue. It wouldn't display leprechaun green, blood red, etc etc.
>
> I hope this is making sense. The visualization challenge is how can
> the taxonomy manager can navigate through this taxonomy and easily
> view all of the relationships and dependencies shared among taxonomy
> facets. The entire data set is quite large.
>
> Sorry for the bad description earlier, I hope this one is better.
>
> Jeff
>
>
> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
> Posted from the new ixda.org
> http://beta.ixda.org/discuss?post=18738
>
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> List Guidelines ............ http://beta.ixda.org/guidelines
> List Help .................. http://beta.ixda.org/help
> Unsubscribe ................ http://beta.ixda.org/unsubscribe
> Questions .................. list at ixda.org
> Home ....................... http://beta.ixda.org
>

--
Kim McGalliard
kimmcg at gmail.com
646-265-8353

27 Jul 2007 - 8:56pm
Roy
2007

There's an example of one approach to a problem very close to this at
http://www.topicscape.com/user-story/mindmaps-for-product-reference.html
This is a case study about how a Jewelry designer's organized both
her own product information and competitors' (to raid for ideas?).
The product images and general information are organized under
colors, styles, brands, stone types, magazines publicized in and so
on.
The software used, Topicscape, shows the information in a 3D
landscape and lets the user (and this particular Topicscape was
designed for use by employees) fly around, going deeper in the
taxonomy according to what they are looking for.

A key element of the software is that one item can appear in many
places without needing copies, so you could focus on color, or focus
on brand and still end up at the right place. In other words it's a
taxonomy or hierarchy that allows multiple parents as well as the
usual multiple children. For this to work, the multi-parented
aspects have to be multiple instances of the same thing, not multiple
copies, or you have a versioning nightmare.

Another thing is that the 3D landscape gives a real "sense of
place" and familiarity makes future searches quicker - you just get
to know your way around.

It does have a word search as well, but as you've indicated the
employee doesn't always know the word to search for. Topicscape
lets them search by concept. The word search itself has a nice
usability feature - the search function is constantly waiting to
pounce: Just type a letter and the search panel appears. No need for
Ctrl F or F3 or other preparation for searching.

Does that come close?
Roy

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://beta.ixda.org/discuss?post=18738

30 Jul 2007 - 2:01pm
White, Jeff
2007

Kim and Roy - those were both very helpful. Thanks!

Jeff

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://beta.ixda.org/discuss?post=18738

Syndicate content Get the feed