Analyzing Personagrams: A Data Portrait Visualization of an Instagram User Activity

Personagram is a visual representation integrated into a web-based tool that displays the entire activity data of one Instagram user. The tool presents two main visualizations of the user data. The first, already evaluated in a previous paper Personal instants, is a more discreet and quantitative representation, in the form of a grid of elements, each one representing one single user activity. The second, Personagram, the one we studied in this paper, is a more abstract and agglomerated way of representing the same data. Here, abstract in the sense that the representation of data is less mensurable or exactly quantified.
Personagram intends to profile the user’s online behavior, as a digital self-portrait. To that extent, the tool offers the possibility of generating a personal and unique artifact from the users’ data, as a form of representing their persona on the social network.
The data used in this work is the data made available for download in the Instagram application itself. It contains information about the user activity such as connections made (green), likes provided (red), published content (dark blue), viewed content (yellow), saved content (purple) and comments made (light blue). It also contains references to the time of day when the different activities were carried out (Figure 1).


Figure 1

Personagrams resulting from one Instagram user’s data.


Analyzing Personagrams


Ten participants conducted the study, five women and five men, with an average of 27 years old, aged between 23 and 30. The participants have an average score of 3 for experience and contact with data visualization, encompassing both inexperienced (score of 2) and experienced participants (score of 4) from different backgrounds, such as computer science, biochemistry, multimedia design and business management.


Figure 2

Personagrams resulting from six different Instagram users’ data.


The work presented itself as a self-portrait of the users’ behavior with the goal of disclosing what kind of user we are online.
A need to readapt the visualizations provided nowadays to us comes from one reflection about our current information society, namely on how it contacts and interacts with its data and its sources of information. In this society, we make qualitative judgments and this nature is relevant to be brought to the visualizations of our personal online data to foster more humane reflections. Thus, we applied a qualitative approach to see if it prompted introspective thoughts and fostered discussions beyond the data itself and not just direct, quantitative deductions.
Personagrams do not need to be fully decoded, that is, do not need to communicate, for example, the exact number for the different types of activities, just need to inform such differences through the elements sizes so that they encourage conversations between people when comparing their outputs and make them think about their similarities and/or differences.
As we can see from the above statements, the more abstract and qualitative component of the Personagrams visual approach has led to more introspective insights, in addition to the typical data insights, when compared with the results of our previous work. We managed to obtain reflections beyond the data and user profiling statements, showing promises in such qualitative approaches.
With the insights gathered, we were able to know where our approach was successful and where it was not. We realized that the mapping of some variables can be improved or reformulated to obtain visual coherence and some redundancies must be rethought or even removed from the visualization as they generated doubts and difficulties in the interpretation of Personagram.
In conclusion, the types of observations provided were promising, meeting our objectives. Their comments reveal subjective and qualitative properties, showing reflections beyond the data and not just comments about data values. The comments are intended to define and understand the authors of the data and not just comment on the differences in the number of different types of activities. Lastly, based on the comments from the participants, we can claim that, while there are still some visual issues to be done, our tool showed promising capabilities of user profiling analysis. As was also suggested in the previous study, participants were interested in the idea of having a personalized physical object with their data, from a simple printed painting to diverse wearables.