Page 1 of 1

We shape our buildings ... how do we know?

Posted: Wed Jul 21, 2010 10:22 am
by Janet Rossini
Seems to me there are lots of topics around the building notion and I thought I'd put this one in a separate thread.

The point of this note is to suggest that we make assumptions about how the space is used, to give an example of an imaginary way to know more about how it's used, and to suggest that knowing would be so useful that we should try to get more information.

In discussions on the forum and standing around in world, people have made suggestions based on assumptions about how our space works. That includes me: I am assuming lots of things:
  • The shape of Kuula is part of why people hang out there, not just "history".
  • The shape of a space strongly influences how it is used.
  • Kuula's Freebies are a bit hard to find, and that is probably good.
  • Ginny's Park is even harder to find and that may not be good.
  • Parts of Kuula, like the news stand, are essentially unused and not paying off well.
Everyone makes observations ... and everyone observes different things. This can make it hard to agree on what is going on, and very hard to agree on what to do.

Just pretend that when you TP into an NCI location, there is some kind of invisible ink on the TP point, that gets on your feet. Suppose that as you walk around, you leave footprints, and that when we turn on the ultraviolet lights, all those footprints show up.

If we took a photo of the footprints, we would see lots of marks where people go often, and fewer where they do not go at all. And suppose that when someone logs or TPs out, they left a little splash in another color. That would be interesting as well.

We know a lot about what that picture would look like ... and there are some areas where we might predict different things.

I think there would be lots of prints right on the TP point, lots of prints on the near stage and near grass, and lots of prints by the freebies. There would be a much smaller number on the home point.

From the home point, people would mostly walk into the small sandbox, a smaller number going toward the park. (You might not agree with this guess.)

From the TP point, there would be a heavy path out to the stage, a less heavy path across the road, and a fairly heavy path from TP to freebies without going out on the stage. (Again, you might not agree.)

We would see a large number of exits from the stage. We would also see a large number from the Freebies. And ... we would see a large number from the TP area, suggesting people coming in, not seeing anything of interest, and leaving.

And so on.

I suggest that footprint photo would be really interesting. We would see patterns, get ideas, and have a common basis for discussion.

We probably can't get that photo. What could we do that would give us more information, like that photo, about how our space is used?

What do we know? How do we know it? What else could we know, and how?

Thanks,

Re: We shape our buildings ... how do we know?

Posted: Mon Nov 01, 2010 2:25 pm
by LindaB Helendale
NCI Dream Seeker in-sim traffic analysis
NCI Dream Seeker in-sim traffic analysis
LindaTrafficAnalyser_collage_DS.jpg (239.68 KiB) Viewed 11318 times
Janet Rossini wrote: Just pretend that when you TP into an NCI location, there is some kind of invisible ink on the TP point, that gets on your feet. Suppose that as you walk around, you leave footprints, and that when we turn on the ultraviolet lights, all those footprints show up.

If we took a photo of the footprints, we would see lots of marks where people go often, and fewer where they do not go at all. And suppose that when someone logs or TPs out, they left a little splash in another color. That would be interesting as well.

We know a lot about what that picture would look like ... and there are some areas where we might predict different things.

I think there would be lots of prints right on the TP point, lots of prints on the near stage and near grass, and lots of prints by the freebies. There would be a much smaller number on the home point.

From the home point, people would mostly walk into the small sandbox, a smaller number going toward the park. (You might not agree with this guess.)

From the TP point, there would be a heavy path out to the stage, a less heavy path across the road, and a fairly heavy path from TP to freebies without going out on the stage. (Again, you might not agree.)

We would see a large number of exits from the stage. We would also see a large number from the Freebies. And ... we would see a large number from the TP area, suggesting people coming in, not seeing anything of interest, and leaving.

And so on.

I suggest that footprint photo would be really interesting. We would see patterns, get ideas, and have a common basis for discussion.

We probably can't get that photo. What could we do that would give us more information, like that photo, about how our space is used?

What do we know? How do we know it? What else could we know, and how?

Thanks,
Sure we can get that photo if we want to :)

I made a low lag av tracker for some other project and had it collect paths when avatars move in kuula and dream seeker. (The tracker produces 0.2 eps on average, so it's like 1/5 of mystitool lag contribution, to assure you I'm not lagging kuula with this.)

The path data is very rich for all kind of analysis, but requires some processing for visualization, you can't just draw 1700 paths over the map and see anything ;)

Below there's some results. It seems I can attach only two images, so I combined some result pics for Kuula and Dream Seeker.

The left image shows the paths avatars move in the sim. The paths are coded with hot steel colormap: most common paths are white, then yelow, red, dark red. The actual paths are wide and have random variation (for example on the road the paths fill the whole road). The images show thinned skeletons of the path distribution and the color shows the relative popularity of the path.

The right images show the distribution of locations where avatars spend time,
with rainbow color coding; red colors show places where most of the time is spent.

Re: We shape our buildings ... how do we know?

Posted: Mon Nov 01, 2010 2:29 pm
by LindaB Helendale
Here's one more image, showing the distribution of visit lengths in Kuula and Dream Seeker.

If you wonder why the curves are not monotonically decreasing, the avatars may move to adjacent sim and come back.

The curves may have a small bias towards too short visits, since I can't have the tracker in the sim over autoreturns, and the tracker rezzes new copy before autoreturn to stay online, but some tracks may get broken during the handover.

Re: We shape our buildings ... how do we know?

Posted: Tue Nov 02, 2010 5:26 pm
by Imnotgoing Sideways
What app are you using for that? Your own work? That is super slick! PHP? I wannit for my site!!! DX

=^-^=

Re: We shape our buildings ... how do we know?

Posted: Tue Nov 02, 2010 9:33 pm
by LindaB Helendale
Imnotgoing Sideways wrote:What app are you using for that? Your own work? That is super slick! PHP? I wannit for my site!!! DX
=^-^=
It's all my own app in lsl and matlab ;) It sends the encoded data to me by email in chunks that fit in mail, and i made a matlab routine to read the data and do the path clustering and image processing. After I polish it a bit to reduce manual work i can certainly do the analysis for your sim Immy.

Re: We shape our buildings ... how do we know?

Posted: Tue Nov 02, 2010 11:15 pm
by Jaco Schaffner
Ooh way advanced compared to my map experiment I have lost the scripting energy for. Very nice and usable data.

Re: We shape our buildings ... how do we know?

Posted: Tue Nov 02, 2010 11:34 pm
by Imnotgoing Sideways
Oh, now I'm all tempted to write a PHP solution. =^-^=

And, Matlab, wise choice. There's an engineer here that uses it for some pretty wicked data analysis presentations. (^_^)

Right now the un-lazy side of me is building a gas station. (^_^)y

Re: We shape our buildings ... how do we know?

Posted: Fri Nov 05, 2010 6:55 pm
by Beverly Montgomery
Interesting!