Brand Interactions on Twitter: A Semantic and Network Graph Approach
-
Project Summary:
-
Data Source:
- Collected approximately 150,000 tweets mentioning @nike, @lululemon, and @adidas from the past 93 days using the Twitter API.
- Tweets are in English and originate from the US, including consumer mentions, corporate messages, and potential spam.
- Analyze Twitter mentions of the three brands through network and semantic analysis to understand user engagement and brand perception.
- Create Twitter Mentions Graph: Build a directed, valued graph to identify central Twitter users for each brand and visualize how mentions flow between users.
- Create Semantic Network Graph: Develop a semantic network to analyze common word associations within tweets for each brand.
- Most Central Users: Identify the top 10 most engaged users for each brand and explore their significance from a marketing perspective.
- Most Important Bridgers: Find users who mention multiple brands and analyze their role in connecting brand communities.
- Inspect Word Clusters: Identify key word clusters and interpret what conversations they represent for each brand.
- Brand-specific Attributes: Determine unique attributes "owned" by each brand through word clusters to reinforce marketing strategies.
- Shared Attributes: Identify common words shared between brands to avoid using non-unique traits in marketing efforts.
