This is a walk-through and discussion of Cyclopath, an open source geo-wiki, a user editable map, that has been up and running in Minneapolis-St Paul for several years. It's interesting because anyone can add data -- points, tags, ratings, notes, even streets -- to the map, and the routing algorithm can immediately take the user-added into account. Cyclopath supports cyclists, but the technology itself is very general and has numerous uses.
Beginners Guide to TikTok for Search - Rachel Pearson - We are Tilt __ Bright...
Tour de cyclopath v10
1. Tour de Cyclopath*
Thomas Erickson
IBM T. J. Watson Research Center
Spring 2010
* Cyclopath is a project of the GroupLens Lab at the University of Minnesota
by Terveen, Preidhorsky, et al. It is open source.
2. Tour de Cyclopath
Why this is of general interest
Cyclopath represents an increasingly important new class of urban application that
• combines human-sourced knowledge with digital data to create a common resource
• provides mechanisms for eliciting and focusing human work to enhance the resource
• enables computations that provide resource-based services
• provides a platform for community collaboration
• and has the potential to serve as a potent symbol of a smarter city
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 2
3. Tour de Cyclopath
Problem: Finding bike-friendly routes around the Twin Cities
Good bike routes differ from good driving routes
5. Although greenway continues in right
direction, take Park Ave due to bike lane
4. This section of bike path goes
through beautiful community gardens
3. Enter greenway bike path
via intersection
2. Take side street that has
lights at two busy crossings
1. Start out in opposite direction to avoid
busy main street
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 3
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
4. Tour de Cyclopath
Problem: Finding bike-friendly routes around the Twin Cities
Good bike routes differ from good driving routes
But much of the information that makes this a good route is not on regular maps
5. Although greenway continues in right
direction, take Park Ave due to bike lane
4. This section of bike path goes
through beautiful community gardens
3. Enter greenway bike path
via “intersection
2. Take side street that has
lights at two busy crossings
1. Start out in opposite direction to avoid
busy main street
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 4
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
5. Tour de Cyclopath
Solution:
Cyclopath
A user-editable map
(a geowiki)
with ‘official’ data (e.g.,
USGS, MNDoT)
and user-entered data
• map objects
• points
• bikeability ratings
• tags
• annotations
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 5
6. Tour de Cyclopath
Solution:
Cyclopath
User editing matters
because ‘official’ data:
• may be missing
• may be incorrect
• may be misaligned
• may need synthesis
• may be dynamic
Furthermore, users can
add data that is
• personal
• timely
• qualitative
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 6
7. Tour de Cyclopath
The User Interface
Map and map key
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 7
8. Tour de Cyclopath
The User Interface
Map and map key
Map controls
• edit, zoom, pan
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 8
9. Tour de Cyclopath
The User Interface
Map and map key
Map controls
Control panels
• request routes, adjust
view, revert changes,
etc.
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 9
10. Tour de Cyclopath
Map elements
Blocks (street)
Points
Block Point
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 10
11. Tour de Cyclopath
Map elements
Blocks (street)
Points
Tags (points)
Notes (points)
Tags
for this point
Notes
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 11
12. Tour de Cyclopath
Map elements
Blocks (street)
Points
Tags (points, blocks)
Notes (points, blocks)
Tags
Notes
for this
block
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 12
13. Tour de Cyclopath
Map elements
Blocks (street)
Points
Tags (points, blocks)
Notes (points, blocks)
Ratings (blocks only)
• personal (private)
• estimated (from others)
• computed Rating
(from MN DoT data)
for this
block
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 13
14. Tour de Cyclopath
Map elements
Blocks (street)
Points
Tags (points, blocks)
Notes (points, blocks)
Ratings (blocks only)
Intersections Intersection
• How streets connect
(or not)
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 14
15. Tour de Cyclopath
Map elements
Blocks (street)
Points
Tags (points, blocks)
Notes (points, blocks)
Ratings (blocks only)
Intersections Intersections?
• How streets connect
(or not)
•Important for
computing routes – data
often missing or
inaccurate for bikes
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 15
16. Tour de Cyclopath
Map elements
Blocks (street)
Points
Tags (points, blocks)
Notes (points, blocks)
Ratings (blocks only)
Intersections
Regions (not shown)
• Public
(neighborhoods)
• Private
(watch regions)
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 16
17. Tour de Cyclopath
Editing
We want users to be
able to edit data
because
• it might be missing
• it might be wrong
• it might be
misaligned
• and users have a
deep qualitative
knowledge of places
the is rarely found in
official data sets
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 17
18. Tour de Cyclopath
Editing example
Here’s a street I added.
I gave it a name, a
type, and a bikeability
rating
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 18
19. Tour de Cyclopath
Editing example
Here’s a street I added.
I gave it a name, a
type, and a bikeability
rating
Later on, someone else
added the tag
“unpaved”
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 19
20. Tour de Cyclopath
Editing example
Here’s a street I added.
I gave it a name, a
type, and a bikeability
rating
Later on, someone else
added the tag
“unpaved”
Later I added a note
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 20
21. Tour de Cyclopath
Editing & reverting
And of course it’s a
wiki so I can
• set a “watch region”
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 21
22. Tour de Cyclopath
Editing & reverting
And of course it’s a
wiki so I can
• set a “watch region”
• and inspect and
revert changes
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 22
23. Tour de Cyclopath
Computing routes
Now we can use all
this data to compute
bike-friendly routes
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 23
24. Tour de Cyclopath
Computing routes
Now we can use all
this data to compute
bike-friendly routes
• Enter From and To
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 24
25. Tour de Cyclopath
Computing routes
Now we can use all
this data to compute
bike-friendly routes
• Enter From and To
• Decide whether to
minimize distance
or favor bikeability
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 25
26. Tour de Cyclopath
Computing routes
Now we can use all
this data to compute
bike-friendly routes
• Enter From and To
• Decide whether to
minimize distance
or favor bikeability
• And select tags to
avoid, bonus or
penalize when
computing route
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 26
27. Tour de Cyclopath
Computing routes
Now we can use all
this data to compute
bike-friendly routes
• Enter From and To
• Decide whether to
minimize distance
or favor bikeability
• And select tags to
avoid, bonus or
penalize when
computing route
Notice that much of
this data is user
entered: point names,
bikeability, tags
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 27
28. Tour de Cyclopath
Computing routes
The route
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 28
29. Tour de Cyclopath
Computing routes
The route
• Can be color-coded
according to various
dimensions (e.g.,
hills, bikeability)
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 29
30. Tour de Cyclopath
Computing routes
The route
• Can be color-coded
according to various
dimensions (e.g.,
hills, bikeability)
• Has a cue sheet
(soon with notes)
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 30
31. Tour de Cyclopath
Computing routes
The route
• Can be color-coded
according to various
dimensions (e.g.,
hills, bikeability)
• Has a cue sheet
(soon with notes)
• Feedback can be
provided
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 31
32. Tour de Cyclopath
Computing routes
The route
• Notice that my route
starts out in the
“wrong” direction –
but that’s really
what I want because
it avoids busy streets
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 32
33. Tour de Cyclopath
Computing routes
The route
• Notice that my route
starts out in the
“wrong” direction –
but that’s really
what I want because
it avoids busy streets
• And it has the other
advantages I
mentioned at the
start of the talk
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 33
34. Tour de Cyclopath
Computing routes
The route
• Notice that my route
starts out in the
“wrong” direction –
but that’s really
what I want because
it avoids busy streets
• And it has the other
advantages I
mentioned at the
start of the talk
• And the route is
also half a mile
shorter than that
offered by Google
Maps’ new bike
routing feature
(even though I
favored bikeability
over distance)
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 34
35. Tour de Cyclopath
Does it work?
Will people really use it?
Will people actually go to the trouble of adding data?
Will the added data make a difference?
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 35
36. Tour de Cyclopath
Does it work?
Usage (in season)
• In use for 1.5 years
• 1,500+ reg. users
• daily: 15-30 reg. &
~150 unreg. users
• 150 routes/day
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 36
37. Tour de Cyclopath
Does it work?
Usage (in season)
• In use for 1.5 years
• 1,500+ reg. users
• daily: 15-30 reg. &
~150 unreg. users
• 150 routes/day
Edits
• ~10,000 edits, by
400+ users
• User input resulted For example: indicating “connectivity” between Como Ave and the
in shorter routes: Intercampus Transitway allowed computation of a new route that
routes 1K shorter is .6 K shorter than the old route
(14.8=>13.8K)
after 9 months
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 37
38. Tour de Cyclopath
Does it work?
Usage (in season)
• In use for 1.5 years
• 1,500+ reg. users
• daily: 15-30 reg. &
~150 unreg. users
• 150 routes/day
Edits
• ~10,000 edits, by
400+ users
• User input resulted Relationship between a person’s views (red) and edits (blue)
in shorter routes:
routes 1K shorter
(14.8=>13.8K)
after 9 months
Individual variability
• Radical variation in
scope of interest and
editing behavior
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 38
39. Tour de Cyclopath
Research Issues:
Eliciting work
Improve map’s routes
Improve others’ routes
Improve your routes
…in progress…
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 39
40. Tour de Cyclopath
Research Issues:
Focusing work
How might Cyclopath
get users to do
particular types of
work in particular
places?
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 40
41. Tour de Cyclopath
Research Issues:
Focusing work
The Work Hints
Cyclopath needs your help
experiment (case 1)
“…We have created a system
which will automatically
direct you to areas of the map
that need work (more
bikeability ratings entered or
edits to the geography of the
map itself)…”
<link to “work hints” window>
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 41
42. Tour de Cyclopath
Research Issues:
Focusing work
The Work Hints
experiment (case 1)
• Direct a person to an
area that needs work
• let them work until
they’re ‘done’
• ask if they want to
do another area
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 42
43. Tour de Cyclopath
Research Issues:
Focusing work
The Work Hints
experiment (case 2)
• Direct a person to an
area that needs work
• let them work until
they’re ‘done’
• ask if they want to
do another area
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 43
44. Tour de Cyclopath
Research Issues:
Focusing work
The Work Hints
experiment
• People did about the
same amount of work
per trial
• BUT they did three
times as many trials:
17.7 trials vs 5.0 trials
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 44
45. Tour de Cyclopath
Research Issues:
Focusing work
The Work Hints
experiment
• People did about the
same amount of work
per trial
• BUT they did three
times as many trials:
17.7 trials vs 5.0 trials
If small maroon circles
can have this kind of
effect, imagine what
could be done by
taking some lessons
from the
ESP game!
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 45
46. Tour de Cyclopath
Research Issues: Summary of Findings
• Visually highlighting work opportunities leads to more work
• Users also do ‘extra’ work (beyond what is visually highlighted_
• Taking users to areas they are familiar with leads to more work of certain types
• Issuing a “call to action” and providing visually highlighting causes a broader range of users to
do work (and moreover the “lead workers” are different)
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 46
47. Tour de Cyclopath
The Future
Cyclopath doesn’t have
to be about bicyclists
• skiers (iceWiki)
• walkers
• disabled
• urban tourists
• local history buffs
• garden clubs
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 47
48. Tour de Cyclopath
The Future
Cyclopath doesn’t have
to be about bicyclists
Nor does it have to be
just for route finding
• Planners
(Cycloplan)
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 48
49. Tour de Cyclopath
The Future
Cyclopath doesn’t have
to be about bicyclists
Nor does it have to be
just for route finding
• Planners
(Cycloplan)
• Energy management
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 49
50. Tour de Cyclopath
The Future
Cyclopath doesn’t have
to be about bicyclists
Nor does it have to be
just for route finding
• Planners
(Cycloplan)
• Energy management
• Resource sharing
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 50
51. Tour de Cyclopath
Takeaways
People have rich and nuanced knowledge of their habitats
People are willing to do work to contribute this knowledge;
furthermore, systems can be designed so that they better elicit and focus such work
If the elicited knowledge is in a form that digital systems can use,
the knowledge can be used in computations and services, as well as being used to refine itself
Why this is of more general importance
Cyclopath represents an increasingly important type of urban application that
• combines human-sourced knowledge with digital data to create a common resource
• provides mechanisms for eliciting and focusing human work to enhance the resource
• enables computations that provide resource-based services
• provides a platform for community collaboration
• and has the potential to serve as a potent symbol of a smarter city
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 51
52. Tour de Cyclopath
The Future
Cyclopath doesn’t have
to be about bicyclists
Nor does it have to be
just for route finding
• Planners
(Cycloplan)
• Energy management
• Resource sharing
• Urban ecology + = ?
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 52
53. Tour de Cyclopath
Why this is of general importance
Cyclopath represents an increasingly important type of urban application that
• combines human-sourced knowledge with digital data to create a common resource
• provides mechanisms for eliciting and focusing human work to enhance the resource
• enables computations that provide resource-based services
• provides a platform for community collaboration
• and has the potential to serve as a potent symbol of a smarter city
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 53
54. End Notes
• Cyclopath was conceived of and developed by the GroupLens Lab at
the University of Minnesota, not IBM (I am just an enthusiast).
• Cyclopath is now open source. Find out more at http://cyclopath.org
5. Although greenway continues in right
direction, take Park Ave due to bike lane
4. This section of bike path goes
through beautiful community gardens
3. Enter greenway bike path
via intersection
2. Take side street that has
lights at two busy crossings
1. Start out in opposite direction to avoid
busy main street
Thomas Erickson, Social Computing Group, IBM T. J. Watson Research Center. slide 54
Terveen, Priedhorsky, et al. ~ GroupLens Lab, EE/CS, University of Minnesota