Problem Domain

Behavioural economics, as the psychology behind economic activities(2), is founded upon the basis of choice(3). Spatial behavior, as a our problem domain, is defined from the perspective of aversiveness and indifference to proximity(4), eg. if opportunities within a constrained spatial range are evaluated only in terms of attraction attributes, then the individual’s spatial choices may be spatially indifferent(4).

However, there exists a variety-seeking tendency, that is an intrinsic tendency to engage a variation in behavior as a goal in and of itself, such variations are known as variety-seeking behavior(5). And yet, spatial indifference is bounded by constraints of willingness to traverse space over a certain spatial range to a set of opportunities that are spatially finite.

Philosophy, Motive, Approach & Hypothesis

Our choice of a philosophy is that: "the intuitive grasp of the essentials of a large complex of facts leads to a basic 'law', from which, as a system of axioms, we may derive conclusions"(7). And as such, even that our chosen domain is foreign to us, we understand that the key to traverse its complex to the clarity of the simple, is studying its ubiquitous language's(8) most essential of constructs, and that is the notion of behavior. On key motive, and with a software engineering background; we address an actor class of our problem domain with a software artifact, to approach our designated notion.

We set our hypothesis as that the economic actor class of spatial choice variety seekers would be better served with a more tailored InfoVis utility other than the popularly used Google Maps.

Project Description

We introduce twalk, an InfoVis utility, intended to serve spatial choice variety seekers in visualizing their spatially finite opportunities from within a specified spatial range defined in terms of origin/destination walking duration.

Project Structure

We follow a simple programming style of single-purpose scripts managed via a Makefile.

Fig: Project Overview

#
# Makefile
#
.POSIX:
.PHONY: all clean help
DATALAKE := ./DATALAKE
DATA := ./DATA
MAP := ./MAP
FIG := ./FIG
$(MAP)/twalk_map.html: $(DATA)/stats.csv ./map.R
@./map.R
## make $(DATA)/stats.csv: creates stats CSV
$(DATA)/stats.csv: $(DATA)/cardinal.csv ./stats.R
@./stats.R
## make $(DATA)/cardinal.csv: creates OD cardinal CSV
$(DATA)/cardinal.csv: $(DATA)/filtered.csv ./cardinal.R
@./cardinal.R
## make $(DATA)/filtered.csv: creates OD filtered CSV
$(DATA)/filtered.csv: ./demo.data.R
@./demo.data.R
## make $(FIG)/vis.makefile.png: shows Makefile's DAG
$(FIG)/vis.makefile.png: ./Makefile
@echo "Visulization of the project's DAG"
@make -Bnd | make2graph | dot -Tpng -o $@
@png2ff < $@ > $(FIG)/vis.makefile.ff
## make help: shows this message
help: Makefile
@sed -n 's/^##//p' $<
## make clean: cleans all
clean:
@rm -rf $(DATA)/*.*
@rm -rf $(MAP)/*.*
@rm -rf $(FIG)/*.*
view raw Makefile hosted with ❤ by GitHub

Data Acquisition

While data acquisition is out of our scope, we needed demo data to demonstrate our project. Our demo data is sampled from Google's Places and MatrixDistance API Endpoints. For simplicity, we maintain a wide schema, until we reach the visualization stage.
e.g. filtered.csv -> cardinal.csv -> stats.csv -> (visualization)

We filter our acquired data by a specified maximum walking duration, as a constraint of our spatial range and store it as filtered.csv, while keeping the original data separately.

o_name o_lat o_lng d_name d_lat d_lng walk_duration walk_duration_value walk_distance walk_distance_value
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Graunas -22.0324573 -47.8804512 56 mins 3353 4.2 km 4238
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Arco Hotel São Carlos by AccorHotels -22.0339686 -47.8797396 56 mins 3378 4.3 km 4270
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Nacional Inn São Carlos -22.0358584 -47.8701406 1 hour 7 mins 4004 5.1 km 5083
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Caiçara Hotel São Carlos -22.0177306 -47.8885453 26 mins 1556 2.1 km 2092
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Anacã -22.009377 -47.890524 13 mins 757 1.0 km 973
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel São Carlos Othon Suítes -22.0079726 -47.899624 7 mins 401 0.6 km 570
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Sleep Inn São Carlos -22.007382 -47.903542 11 mins 652 1.0 km 959
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Ypê de São Carlos LTDA -22.01134 -47.890997 15 mins 882 1.1 km 1139
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Parisi Hotel São Carlos -22.0051698 -47.8911179 10 mins 581 0.8 km 820
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 ibis Sao Carlos -22.0179347 -47.9179691 41 mins 2447 3.3 km 3256
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Central Park Hotel - São Carlos -22.00118 -47.899929 14 mins 859 1.2 km 1245
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Accacio -22.0158656 -47.8909126 21 mins 1241 1.6 km 1650
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Apart Suites -22.0180233 -47.8943027 36 mins 2133 2.8 km 2813
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Indaiá Hotel Residence -22.0060419 -47.8904349 10 mins 607 0.8 km 818
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Residence -22.0044424 -47.8959255 6 mins 379 0.6 km 570
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Executive -21.9875836 -47.8950971 36 mins 2133 2.8 km 2813
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotéis em Sao Carlos -Hotel, Pousada ,Hostel -22.0115131 -47.890697 15 mins 882 1.1 km 1139
view raw filtered.csv hosted with ❤ by GitHub

Cardinal Direction Ranking

First, we used our origin/destinations data to calculate to which cardinal direction each destination belongs. We used R's GeoSphere bearingRhumb() function. And we relaxed the cardinal accuracy of the destinations to provide a useful abstraction of a walking direction choice.
o_name o_lat o_lng d_name d_lat d_lng walk_duration walk_duration_value walk_distance walk_distance_value cardinal
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Graunas -22.0324573 -47.8804512 56 mins 3353 4.2 km 4238 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Arco Hotel São Carlos by AccorHotels -22.0339686 -47.8797396 56 mins 3378 4.3 km 4270 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Nacional Inn São Carlos -22.0358584 -47.8701406 1 hour 7 mins 4004 5.1 km 5083 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Caiçara Hotel São Carlos -22.0177306 -47.8885453 26 mins 1556 2.1 km 2092 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Anacã -22.009377 -47.890524 13 mins 757 1.0 km 973 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel São Carlos Othon Suítes -22.0079726 -47.899624 7 mins 401 0.6 km 570 W
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Sleep Inn São Carlos -22.007382 -47.903542 11 mins 652 1.0 km 959 W
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Ypê de São Carlos LTDA -22.01134 -47.890997 15 mins 882 1.1 km 1139 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Parisi Hotel São Carlos -22.0051698 -47.8911179 10 mins 581 0.8 km 820 NE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 ibis Sao Carlos -22.0179347 -47.9179691 41 mins 2447 3.3 km 3256 SW
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Central Park Hotel - São Carlos -22.00118 -47.899929 14 mins 859 1.2 km 1245 NW
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Accacio -22.0158656 -47.8909126 21 mins 1241 1.6 km 1650 SE
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Apart Suites -22.0180233 -47.8943027 36 mins 2133 2.8 km 2813 S
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Indaiá Hotel Residence -22.0060419 -47.8904349 10 mins 607 0.8 km 818 E
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Residence -22.0044424 -47.8959255 6 mins 379 0.6 km 570 N
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Executive -21.9875836 -47.8950971 36 mins 2133 2.8 km 2813 N
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotéis em Sao Carlos -Hotel, Pousada ,Hostel -22.0115131 -47.890697 15 mins 882 1.1 km 1139 SE
view raw cardinal.csv hosted with ❤ by GitHub
We calculate our descriptive statistics of each cardinal e.g. standard deviations, means, and coefficient of variations; as for each cardinal direction we may represent walking duration spread via the coefficient of variations, spatial proximity via the mean and spatial variety via the destination count.

However, we choose to scope our current work only to the use of destination counts within a specified spatial range at each direction, as a guiding score for spatial variety.

o_name o_lat o_lng d_name d_lat d_lng walk_duration walk_duration_value walk_distance walk_distance_value cardinal cardinal_sd cardinal_mean cardinal_cv min_cv cardinal_count rank scale
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Parisi Hotel São Carlos -22.0051698 -47.8911179 10 mins 581 0.8 km 820 NE NA 581 NA NA 1 1 45.1721170395869
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Indaiá Hotel Residence -22.0060419 -47.8904349 10 mins 607 0.8 km 818 E NA 607 NA NA 1 1 43.2372322899506
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Apart Suites -22.0180233 -47.8943027 36 mins 2133 2.8 km 2813 S NA 2133 NA NA 1 1 12.3042662916081
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 ibis Sao Carlos -22.0179347 -47.9179691 41 mins 2447 3.3 km 3256 SW NA 2447 NA NA 1 1 10.7253780138946
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Central Park Hotel - São Carlos -22.00118 -47.899929 14 mins 859 1.2 km 1245 NW NA 859 NA NA 1 1 30.5529685681024
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Residence -22.0044424 -47.8959255 6 mins 379 0.6 km 570 N 1240.2652942012 1256 0.987472367994589 FALSE 2 2 69.2480211081794
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 The Hill Hoteis Executive -21.9875836 -47.8950971 36 mins 2133 2.8 km 2813 N 1240.2652942012 1256 0.987472367994589 FALSE 2 2 12.3042662916081
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel São Carlos Othon Suítes -22.0079726 -47.899624 7 mins 401 0.6 km 570 W 177.483802077823 526.5 0.337101238514385 TRUE 2 2 65.4488778054863
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Sleep Inn São Carlos -22.007382 -47.903542 11 mins 652 1.0 km 959 W 177.483802077823 526.5 0.337101238514385 TRUE 2 2 40.2530674846626
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Graunas -22.0324573 -47.8804512 56 mins 3353 4.2 km 4238 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 7.82731881896809
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Arco Hotel São Carlos by AccorHotels -22.0339686 -47.8797396 56 mins 3378 4.3 km 4270 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 7.76939017169923
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Nacional Inn São Carlos -22.0358584 -47.8701406 1 hour 7 mins 4004 5.1 km 5083 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 6.5546953046953
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Caiçara Hotel São Carlos -22.0177306 -47.8885453 26 mins 1556 2.1 km 2092 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 16.8669665809769
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Anacã -22.009377 -47.890524 13 mins 757 1.0 km 973 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 34.669749009247
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Ypê de São Carlos LTDA -22.01134 -47.890997 15 mins 882 1.1 km 1139 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 29.7562358276644
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotel Accacio -22.0158656 -47.8909126 21 mins 1241 1.6 km 1650 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 21.1482675261886
Institute of Mathematics and Computer Sciences -22.0070347 -47.8949323 Hotéis em Sao Carlos -Hotel, Pousada ,Hostel -22.0115131 -47.890697 15 mins 882 1.1 km 1139 SE 1339.68396459762 2006.625 0.667630456411946 FALSE 8 8 29.7562358276644
view raw stats.csv hosted with ❤ by GitHub

Information Visualization

Data is insufficient for meaningful communication while information, as data transformed and interpreted within context is. And while Visualization is a cognitive activity, facilitated by external stimuli from which we build internal mental representations of the world (1); Information Visualization is the process that provides a capability, for transforming data into information(1).

We note that the more specific the question we are asking, the more specific and clear the visual result will be(6). And as such our single-purpose utility employs minimal visual constructs to deliver its intended answer to the question:

Which walking direction(s) would offer more opportunities for a spatial choice variety seeker?

And to answer, our process of transforming data into information included the feature use of spatial orientation, color, size, and user interaction:
1. We provided our users with a spatial orientation via the use of a map.
2. We offered a map's legend to color associate a spatial direction with destination counts.
3. We represented walk duration proximity via size, the bigger the radius of a destination circle, the closer it is to origin.
4. And finally, we offered information on-demand via user's interaction e.g. popup interaction.


Conclusion

Spatial indifference is bounded by constraints of willingness to traverse space over a certain range, and regardless of variety-seeking behavior, the set of opportunities considered by an individual when making a spatial choice are spatially finite. We employed cardinal directions and destination counts to offer, through spatial visual constructs of color and size, each walking direction opportunities. And as we are spatial choice variety seekers ourselves, we find that our utility yeilds, more than Google Maps, a visually instant answer to the question: Which walking direction(s) would offer more opportunities for a spatial choice variety seeker? and hence we accept our hypothesis. In aggregate, Information Visualization offers a much needed aid, and our introduced utility presents useful visualization of the spatially finite opportunities for its intended audience.

Acknowledgment

We are very grateful to Osafu Augustine Egbon (NUSP: 11494792), an ICMC Statistics PostGrad, for his enthusiasm and diligence in fulfilling our various statistical inquiries.

Resources

1. Places API,
2. Distance Matrix API
3. R
4. GeoSphere
5. Make

Please refer to our github page for complete sourcecode and collected demo data.

References

(1) "Introduction to Information Visualization", Riccardo Mazza, Springer, 2009
(2) "Theory of Games and Economic Behavior", Oskar Morgenstern and John von Neumann, Princeton University Press, 1944
(3) "Urban Dynamics and Spatial Choice Behaviour", Joost Hauer, Harry Timmermans, and Neil Wrigley, Kluwer Academic Publishers, 1989
(4) "Spatial Choice Theory and Spatial Indifference: A Comment", Frank Stetzer and Alan G. Phipps
(5) "Variety-seeking in Product Choice Behavior", Hans (J.C.M.) van Trijp, Pudoc Scientific Publishers 1995
(6) "Visualizing Data", Ben Fry, O'Reilly, 2008
(7) "Building Theories: Heuristics and Hypotheses in Sciences", David Danks and Emiliano Ippoliti, Springer, 2018
(8) "Domain-driven Design: Tackling Complexity in the Heart of Software", Eric Evans, Addison Wesley, 2003