This project applies Multidimensional Scaling (MDS) to analyze and group social media influencers based on various characteristics and metrics such as engagement rate, follower count, content type, and audience demographics. The goal is to visualize the influencers in a lower-dimensional space, making it easier to identify patterns, similarities, and groupings among them.
So, subjectively, new groups are formed that show high similarity between Tik Tok social media influencers.
-
TikTokers named therock, belindatok, and landonbarkerr have similarities based on the five aspects so they are put into the same group.
-
TikToker Named mrbeast and onwardwanna are included in the same group because they have similarities based on all five aspects.
-
TikToker Named addisonre, khaby.lame, scarlettsspam2, jypestraykids, and bizzarrap do not have similarities in all five aspects so they cannot be included in the same group.
This shows their position that is not close to any influencer on the two-dimensional map, which means that based on the criteria measured, they do not have similarities with other influencers.
Translated with DeepL.com (free version)
- Data Processing: Preprocessing of influencer data including normalization and transformation for MDS compatibility.
- MDS Visualization: Visual representation of influencers in 2D/3D space based on their similarities.
- Grouping and Clustering: Analysis of the resulting MDS plots to identify natural groupings or clusters of influencers.
- Insights: Provides insights into influencer strategies and content categories that appeal to different audience segments.
- R Language
- R Studio
- Multidimensional Scaling (MDS) for dimensionality reduction
- Brand Marketing: Helps brands identify influencers who share similar audience profiles.
- Campaign Strategy: Assists in selecting diverse influencer groups to maximize campaign reach and impact.