Report Link
https://selfiecity.net/#findings
Visualisation Link
https://selfiecity.net/#
Visualisation Summary
The visualisation is the result of a study done with the support of The Graduate Center, City University of New York, California Institute for Telecommunication and Information, and The Andrew W. Mellon Foundation. The study aimed at doing a demographic study of people taking selfies especially in terms of poses and expressions. The observations are represented as imageplots.
Visualisation Critique
Tools:
The visualisation used Instagram to gather the data and 120,000 photos from different cities were selected. Help was used from Amazon Mechanical turk to tag these pictures. 1000 pictures selected from each city were reduced to 640 and they were run through facial analysis software for tagging. Image plots and combination of inferences from these were plotted against various visualisations.
What works:
- The imageplots clearly show the number of males and females across age ranges clicking selfies with age on x-axis and no of selfies on y. Gender is restricted to a specific quadrant.
- The second set of image plots use the same method but now plot no-of selfies on mood. It is again an easy to consume visualisation.
- The analysis plot helps user customise each factor like age, turn, tilt of head etc view the number of selfies taken. The area graph is an appropriate use case her for lot of the sub-factors
What doesn't work:
- Imageplot make it difficult to capture more than 3 factors eg: age, number of selfies , gender and mood would be a challenge to plot on the same graph
- The analysis section uses circular charts for things like gender and demography and it does not clearly represent the numbers that the chart is trying to represent. The data seems just visual with no statistics or number behind it for these specific factors.