We present a method by which to convert a large corpus of website traffic data into interactive and practical taste graph visualizations. The website traffic data lists individual visitors' level of interest in specific pages across the website; it is a tripartite list consisting of anonymized visitor ID, webpage ID, and a score that quantifies interest level. Taste graph visualizations reveal psychological profiles by revealing connections between consumer tastes; for example, an individual with a taste for A may be also have a taste for B. We describe here the method by which we map the web traffic data into a form that can be displayed as interactive taste graphs, and we describe design strategies for communicating the revealed information. In the context of the publishing industry, this interactive visualization is a tool that renders the large corpus of website traffic data into a form that is actionable for marketers and advertising professionals. It could equally be used as a method to personalize services in the domains of government services, education or health and wellness.
Citation: |
Figure 2. Once a category is selected, the user enters the main visualization view and is prompted to define a target group. The menu on the upper far left allows the user to navigate to a different category if s/he wants to browse through all the interests (website pages). A group is defined as all individuals with a common interest (or individuals who have visited a common page); the user choses the interest from a display on the left side of the visualization. The size of the bubbles in the group display on the left is proportional to the total number of visitors for each of the pages. The taste display on the right remains blank until the user has defined a group.
Figure 3. Here the user selects American Food in from the group display on the left. This defines the target group as all the individuals who have looked at (shown an interest in) American Food recipes in the website. Once the target group has been defined, the taste display on the right is populated. The size of the bubbles in the taste display is proportional to the number of target group members that have visited each of the pages. Above the taste display, there is a prompt for the user to select one or more tastes.
Figure 4. The user can use the menu on the far right to navigate to a different category. Here, the user has navigated the taste category from Food and Drink in figure 3 to Hobbies and Interests shown here. The user is still prompted to select one or more tastes.
Figure 5. Here, there user has selected Video Games and Ceramics. This opens a window that allows the user to compare the target group's degree of interest in the selected tastes. The window has two tabs. The Raw Scores tab, shown here, displays the selected group's interest score for Video Games and Ceramics, as well as the reference group's interest scores. The default reference group is general public, which comprises all individuals in the data set.
Figure 7. If the user wants the reference group to be something other than the 'general public', then s/he can define a reference group by selecting another interest in the group display on the left. Here the user selects Asian Food from the display. This defines the reference group as all the individuals who have looked at (shown an interest in) Asian Food recipes in the website. This view shows the detail window open to the Raw Scores tab displaying each group's interest scores in the selected tastes.
[1] |
A. Cairo,
The Functional Art: An Introduction to Information Graphics and Visualization New Riders, Berkeley, CA, 2013.
![]() |
[2] |
M. Bostock,
D3. js -Data-Driven Documents Available from: https://d3js.org/.
![]() |
[3] |
L. Kozlowski, Gravity can graph your internet clicks for new customer snapshots,
http://www.forbes.com/sites/lorikozlowski/2013/11/13/brand-graphs-a-new-snapshot-of-consumers/#7adbf713a2d1, forbes. com 11. 13. 2013. Web 6. 1. 2016.
![]() |
[4] |
I. Krumpal, Determinants of social desirability bias in sensitive surveys: A literature review, Qual. Quant., 47 (2013), 2025-2047.
doi: 10.1007/s11135-011-9640-9.![]() ![]() |
[5] |
V. Kumar, Building a Taste Graph: The basic principles,
http://bigdata-madesimple.com/category/tech-and-tools/analytics/, bigdata-madesimple. com 12. 23. 2014. Web 6. 1. 2016.
![]() |
[6] |
S. Pearman, Delicious interest graphs: Taco bell and whole foods,
http://www.gravity.com/blog/delicious-interest-graphs-taco-bell-and-whole-foods/, Gravity. com 8. 12. 2013. Web 6. 1. 2016.
![]() |
[7] |
S. Pearman, What Your Electric Car Says About You,
http://www.gravity.com/blog/what-your-electric-car-says-about-you/, Gravity. com 7. 31. 2013. Web 6. 15. 2016.
![]() |
[8] |
B. Shneiderman, The eyes have it: A task by data type taxonomy for information visualizations, IEEE Symposium on Visual Languages Proceedings, (1996), 336-343.
doi: 10.1109/VL.1996.545307.![]() ![]() |
[9] |
A. Taylor, Snow Fight: Skiing versus Snowboarding,
http://www.gravity.com/blog/snow-fight-skiing-versus-snowboarding/, Gravity. com 2. 17. 2015. Web 6. 1. 2016.
![]() |
[10] |
E. R. Tufte,
The Visual Display of Quantitative Information 2nd edition, Graphics Press, Cheshire, Conn., 2001.
![]() |
[11] |
C. Ware,
Information Visualization: Perception for Design Elsevier, Waltham, MA, 2013.
![]() |
[12] |
N. Yau,
Visualize This John Wiley & Sons, Indianapolis, Indiana, 2011.
doi: 10.1002/9781118722213.![]() ![]() |
[13] |
J. Yu and H. Cooper, A quantitative review of research design effects on response rates to questionnaires, J. Mark. Res., 20 (1983), 36-44.
doi: 10.2307/3151410.![]() ![]() |
[14] |
T. Zhou, J. Ren, M. Medo and Y. -C. Zhang, Bipartite network projection and personal recommendation Phys. Rev. E 76 (2007), 046115.
doi: 10.1103/PhysRevE.76.046115.![]() ![]() |