Social media app-preference of young adults by personality profile

Hello there!

In this project, you are tasked with generating a spider-web diagram on the relation between social-media applications use and the personality profiles of more than 3000 users. We want to focus on the data regarding “young adults” (18 to 34) in particular. The social media platforms included are Whatsapp, Facebook and Instagram, and the personality profiles are are based the “Big-5” trait-model of personality, which measures Openness, Conscientiousness, Extraversion, Agreeableness and Neuroticism by means of a standardized questionnaire.

Data

The data can be found under the “suplementary material” of this paper, on the right side of the page. The “Data sheet 1” contains an explanation of the variables in “Data sheet 2”; this is where all the info really is. Here you will find the variables: Age Gender, Education, has_whatsapp, has_facebook and has_instagram, followed by 45 variables encoding for the personality questionnaire answers. Luckily, we can ignore this and focus on the personality trait final scores, encoded as BFI_Extra, BFI_Agree, BFI_Consc, BFI_Neuro and BFI_Open. We can also ignore the subdivisions (e.g. BFI_Open_Aesth, BFI_Open_Idea).

Tables 1 through 5 present important summary statistics that relate social media use to age (1), gender (2), education (3) and personality profile (4-5), respectively, but the format cannot be directly worked with, so any valuable info contained here has to be recreated form “Data sheet 2”.

Data manipulation

We want to focus on the 18 to 34 age group (arbitrarily), so you will have to filter the data accordingly. We also want to disregard the differences in gender and education level to look only at those associated with personality, which means you will have to average across those variables.

We are interested in the personality-trait values of users of single apps (W, F, I), but also of users of multiple apps or of none (W-F, F-I, W-I, W-F-I, none). This gives us 8 different combinations (represented for instance in Table 5). The aim is to plot a different “web” in the diagram for each of these possibilities, so you might want to create separate columns to store the values for the combinations, which are not explicitly included in “Data sheet 2” (tip: the variables has_whatsapp, has_facebook, has_instagram encode the info of each user’s app use).

Data visualization

The goal is to create a spider-web diagram that looks like the one in the example below. In a such a diagram, the distance from the center represents the average score of the personality trait, and the “angle” represents the trait itself. You should include a title and a legend. The personality components should be labeled in five points on the margin of the diagram. The filling of the different social media platforms should be semitransparent and of different colors, so that the overlaps and the differences between the “webs” are more apparent. Ideally, labeled concentric circles (0.2, 4.0 …) should also be included to show the degree of association with each personality trait.

Note the illustration below includes Twitter, snapchat and LinkedIn, which are not part of the data. Instead, the legend should include the 8 combinations of social media use mentioned above.

Which conclusions/hypotyhesis can we draw from the resulting graph? Do they match some of the resluts of the study (found in “Presentation 1”)?

Best of luck on your implementation!