Visualization

[expatrick]

Hello,
found the group via Selena's tweet. I'm excited, I have been wanting
to learn more about R since the Times article came out. I deal with
visualization issues in my job at RNA networks, where I built our
product's GUI and maintain marketing dashboards. I'm a coder, MBA, and
artist and have several projects that could benefit from better
visualization and feedback.

I'm looking forward to meeting you and learning what you're doing. In
the meantime, have you all seen these amazing maps representing made
from Craig's List Missed Connections posts?
http://www.verysmallarray.com/?p=774

Patrick

[Ragav Satish]

> I'm looking forward to meeting you and learning what you're doing. In
> the meantime, have you all seen these amazing maps representing made
> from Craig's List Missed Connections posts?
> http://www.verysmallarray.com/?p=774
>
> Patrick

Patrick,

Do you know if this is program generated or hand drawn? It was my
impression that form fitting varied words on arbitrary shapes(like a
map) was a very hard problem .. so I'm curious on how this was done.

If its hand drawn, then well..

--Cheers
--Ragav

[Patrick Sullivan]

Pretty sure its hand drawn. And well - that's awesome.

-----Original Message-----
From: Ragav Satish <ragav...@gmail.com>

Date: Mon, 02 Mar 2009 23:02:59
To: <pdx-visu...@googlegroups.com>
Subject: Re: Visualization

[Ragav Satish]

On Tue, 2009-03-03 at 07:35 +0000, Patrick Sullivan wrote:
> Pretty sure its hand drawn. And well - that's awesome.

Thanks, didn't mean to imply that it wasn't awesome ;-). But the
awesomeness would have increased by a few orders of magnitude if it was
generated.

--Cheers
--Ragav

[Patrick Sullivan]

That would be cool. I think it would be tough to make it look that good. In my opinion, the best visualizations have been done by hand. Obviously that's not practical for ongoing monitoring types of visualizations, but worth the effort for a visualization that's intended to persuade.

The Craig's list visualization really serves no purpose, which may be why it's so appealing to me. I am really excited to meet you all and find out what tools are out there. Thank you for setting this up.

Patrick

[Amber Case]

Patrick,

I'd really be interested in talking to you more about marketing
dashboards, as these are things I create do often. Aside from that, I
look forward to meeting you in general. I hope you can make it to the
WebTrends meetup, and say hello to Selena for me! Re: Very small array
» Missed Connections: Where, Exactly - very interesting -- viz I had
not seen before.

Ed, R might be really great to go into. I'm interested in it too.
We'll see where the meeting takes us. I'm sure we'll all have a lot to
discuss.

-Amber / caseorganic

[Patrick Sullivan]

Amber, now that I see your blog, I think we met at Urban Grind NW. I walked up to your table and butted into your conversation and mentioned a bit about http://lov.li

I once created some visualizations of the lov.li community with graphviz that were really beautiful. I'll see if I can find them for the meetup.

Patrick

[M. Edward (Ed) Borasky]

On Tue, Mar 3, 2009 at 9:48 AM, Patrick Sullivan <expa...@gmail.com> wrote:
> Amber, now that I see your blog, I think we met at Urban Grind NW. I walked
> up to your table and butted into your conversation and mentioned a bit about
> http://lov.li
>
> I once created some visualizations of the lov.li community with graphviz
> that were really beautiful. I'll see if I can find them for the meetup.

Patrick, are you a GraphViz guru? As powerful as it is and as long as
it's been around, it's surprising to me that it isn't better known.
And you can get to GraphViz from just about every programming language
there is. Would you be willing to do a GraphViz demo / getting started
session?
--
M. Edward (Ed) Borasky
http://www.linkedin.com/in/edborasky

I've never met a happy clam. In fact, most of them were pretty steamed.

[Patrick Sullivan]

Haha! I'm not a guru but it's super easy to generate some nice looking images. I'd be glad to show how it works.
Patrick

[Amber Case]

Patrick, 

That is fantastic news. I remember meeting you very briefly as well. I showed http://lov.li at CyborgCamp and would be very interested in seeing how you used graphviz to generate visualization of it. I'd love to see a short GraphViz demo. 

-Amber 

[Amber Case]

I just heard back from Gregg Webber, Facilities Manager at WebTrends. 

He says that the Half Dome can accommodate about 50 people, the lunch room is currently

set up with a seating capacity of 85 but we could add chairs to bring that over 100 if necessary.

Very quiet space - we can bring beer/food if we want, ect.  

-Amber 

[Micah Elliott]

On Tue, Mar 3, 2009 at 10:01 AM, Patrick Sullivan <expa...@gmail.com> wrote:
> Haha! I'm not a guru but it's super easy to generate some nice looking
> images. I'd be glad to show how it works.

I'm not a graphviz guru either, but have messed with it a bit in
the past. I've meant to make use of the MoinMoin plugin, as seen
on some Mercurial wiki pages
(http://www.selenic.com/mercurial/wiki/index.cgi/UnderstandingMercurial).

A few simple "getting started" resources for those who want a head start:
* apt-get install graphviz python-pydot python-pygraphviz
(Python/Ubuntu friendly)
* man graphviz
* tutorial: http://cli.gs/1BS6nn
* some flickr example images, some with code: http://cli.gs/4g8bPt
* simplest use: dot hello.dot -Tpng -o hello.png && firefox hello.png
* use from Python: http://code.google.com/p/pydot/

--
@MicahElliott | m...@MicahElliott.com | http://MicahElliott.com
Sent from: Beaverton OR United States.

[M. Edward (Ed) Borasky]

Patrick Sullivan wrote:
> Haha! I'm not a guru but it's super easy to generate some nice looking
> images. I'd be glad to show how it works.
> Patrick
>
>
> On Tue, Mar 3, 2009 at 9:53 AM, M. Edward (Ed) Borasky <zzn...@gmail.com
> <mailto:zzn...@gmail.com>> wrote:
>
>
> On Tue, Mar 3, 2009 at 9:48 AM, Patrick Sullivan
> <expa...@gmail.com <mailto:expa...@gmail.com>> wrote:
> > Amber, now that I see your blog, I think we met at Urban Grind NW.
> I walked
> > up to your table and butted into your conversation and mentioned a
> bit about
> > http://lov.li
> >

> > I once created some visualizations of the lov.li <http://lov.li>

> community with graphviz
> > that were really beautiful. I'll see if I can find them for the
> meetup.
>
> Patrick, are you a GraphViz guru? As powerful as it is and as long as
> it's been around, it's surprising to me that it isn't better known.
> And you can get to GraphViz from just about every programming language
> there is. Would you be willing to do a GraphViz demo / getting started
> session?
> --
> M. Edward (Ed) Borasky
> http://www.linkedin.com/in/edborasky
>
> I've never met a happy clam. In fact, most of them were pretty steamed.
>
>
>
>
> >

Excellent!!

P.S.: R has a library that can plot graphs via GraphViz and more
libraries for manipulating graphs as either Node-Edge Lists or adjacency
matrices. Apologies for geeking out :)

http://www.bioconductor.org/packages/release/bioc/html/Rgraphviz.html
http://bioconductor.org/packages/2.3/bioc/html/graph.html

[M. Edward (Ed) Borasky]

In the computer capacity planning and performance monitoring world,
there is a rather nifty tool called "moodss", which stands for
"Modular Object-Oriented Dynamic Spreadsheet". I haven't done much
with it, since I love building my own tools, but it's very popular.
There are some screen shots at

http://moodss.sourceforge.net/

[Amber Case]

Ed,

This is very interesting. My friends and I have figured out a way to
make dynamically updating Google spreadsheets, but nothing in terms of
Micro Excel spreadsheets. Would be interesting to see it.

-Amber

[Amber Case]

Ahh -- is for Linux. I'm looking at bridging gap between tech and
accessibility. There's one way which is to go very beautiful but lose
interaction and meaning -- and the other way to get technical but
difficult to use.

-Amber

[Thomas Lockney]

On Tue, Mar 3, 2009 at 3:25 PM, Amber Case <caseo...@gmail.com> wrote:
>
> Ed,
>
> This is very interesting. My friends and I have figured out a way to
> make dynamically updating Google spreadsheets, but nothing in terms of
> Micro Excel spreadsheets. Would be interesting to see it.

Of course, you can use the Google APIs to update Google speadsheets:
http://www.mattcutts.com/blog/write-google-spreadsheet-from-python/

I've seen a few similar posts/docs in the past about using APIs from
various languages for updating MS Excel docs. I try not to get too
close to Excel these days, if I can help it, but I'm sure I could dig
up some links if you are interested.

[M. Edward (Ed) Borasky]

On Tue, Mar 3, 2009 at 3:33 PM, Amber Case <caseo...@gmail.com> wrote:
>
> Ahh -- is for Linux. I'm looking at bridging gap between tech and
> accessibility. There's one way which is to go very beautiful but lose
> interaction and meaning -- and the other way to get technical but
> difficult to use.

The base capabilities of moodss are for Linux capacity planning. But
it's modular, so if you can get the data into it from anywhere, you
can make dashboards from the data.

Of course, since I'm highly technical, I find beauty in scatterplot
matrices, grand tours, spin, stop and brush, etc. :) One of my friends
referred to GGobi as his favorite video game.

[Amber Case]

I'd like to get more technical and understand what I'm seeing on a
more granular level. Of course, I'm still coming from a background in
anthropology - so I do a lot of research on the structure and function
of complex networks. This has to do with a lot of graphs and seeing
how information can be useful to many groups of people.

[M. Edward (Ed) Borasky]

If you hang out with R programmers as much as I do, you'll find that
they/we think R, which is a programming language first and a graphics
/ statistics package second, did things the right way and everybody
else -- Excel, SPSS, SAS, Minitab -- did it wrong. The others started
with an intuitive GUI and tacked on the programming language as an
afterthought because the users didn't want to do all that typing, but
wanted to build applications.

But of course, R also got beat up because it *didn't* have an
"intuitive GUI", so they built Rpad and R Commander:

http://cran.r-project.org/web/packages/Rpad/index.html
http://cran.r-project.org/web/packages/Rcmdr/index.html

So we end up in the same place -- we just take different paths. :)

[Patrick Sullivan]

coincidentally, I was just demoing google forms -> google spreadsheets.

I just got done with a meeting for a pretty interesting initiative - setting up a very simple, very repeatable direct-to-consumer sales model for tribal fishermen so they can capture the retail markup instead of just the wholesale. For a proof of concept, we want to get a few fishermen taking pre-orders online and showing up at farmer's markets to sell the fish. We're talking about a 2 page site with an order form that is an embedded google form. Unfortunately, these embedded forms are ugly. They live in an iframe, and it appears that in order to dress them up with CSS you have to break them out of the iframe and trick google a bit. The other option would be to create a simple web frontend to the spreadsheet. Then you have to create the validation functions, and the big problem is that the complexity of a custom solution creates a dependency on someone like me, who has plenty of time now, but not so much when I start law school in the fall. This is a perennial problem for non-profits.

Also, these fishermen are not very web-savvy (although they are SMS pros). Their kids are, though.

Anyway, the point is the same - trading off between beauty and simplicity, free stuff and custom software, independence from coders and rich functionality.

Not a visualization conversation anymore I guess. but interesting I hope.

P

[M. Edward (Ed) Borasky]

On Tue, Mar 3, 2009 at 4:29 PM, Patrick Sullivan <expa...@gmail.com> wrote:
> coincidentally, I was just demoing google forms -> google spreadsheets.
>
> I just got done with a meeting for a pretty interesting initiative - setting
> up a very simple, very repeatable direct-to-consumer sales model for tribal
> fishermen so they can capture the retail markup instead of just the
> wholesale. For a proof of concept, we want to get a few fishermen taking
> pre-orders online and showing up at farmer's markets to sell the fish. We're
> talking about a 2 page site with an order form that is an embedded google
> form. Unfortunately, these embedded forms are ugly. They live in an iframe,
> and it appears that in order to dress them up with CSS you have to break
> them out of the iframe and trick google a bit. The other option would be to
> create a simple web frontend to the spreadsheet. Then you have to create the
> validation functions, and the big problem is that the complexity of a custom
> solution creates a dependency on someone like me, who has plenty of time
> now, but not so much when I start law school in the fall. This is a
> perennial problem for non-profits.
>
> Also, these fishermen are not very web-savvy (although they are SMS pros).
> Their kids are, though.
>
> Anyway, the point is the same - trading off between beauty and simplicity,
> free stuff and custom software, independence from coders and rich
> functionality.
>
> Not a visualization conversation anymore I guess. but interesting I hope.

Yeah ... I'd have to do some searching, but I recall reading an
article a few months ago about some ocean fishermen somewhere in Asia
-- Bangladesh, I think. It turned out of all the "advanced"
technologies they could use, the one that was most useful was "plain
old ordinary cell phones." And I'm pretty sure it was voice -- not
text -- that was their primary modus operandi.

Amber, have cell phones really become "plain old ordinary cell phones?" :)

[Amber Case]

Ah - very interesting. Now I am interested in R more than ever.

[Amber Case]

Re: Plain old ordinary cell phones - yes b/c of the rapid creation of
value in a liquid economy with less friction than ever before.

[Bram Pitoyo]

Talking about all these reminded me that, recently, I’ve found out
ways to tie more data from different channels together. This may be
closer to data plumbing than data visualization, though.

[Amber Case]

I'd like to see this demonstrated. I think that data plumbing and data
vizr very closely tied together. The edges between them, when blurred,
provide a lot of usefullness and insights.