About This Talk
At the Academy of Analytics 2020, I combined social network theory mathematics with practical data visualization — demonstrating the Friendship Paradox and its implications through Tableau, while also covering the DWH department’s role and architecture at Mail.ru Group.
Key Ideas
The Friendship Paradox — A counter-intuitive mathematical property of networks: on average, your friends have more friends than you do. This isn’t a subjective feeling — it’s a provable statistical phenomenon that arises from the structure of social networks. High-degree nodes (people with many connections) appear disproportionately in others’ friend lists.
Implications for Business — The friendship paradox has practical applications in viral marketing (target friends of random users for faster spread), influence measurement (degree centrality overestimates the “average” experience), epidemic modeling, and network sampling strategies.
Visualization in Tableau — The talk demonstrated techniques for visualizing network data in Tableau: converting adjacency lists to coordinate layouts, using calculated fields for node sizing based on centrality metrics, and creating interactive network explorations. Tableau isn’t a natural network visualization tool, but with the right data preparation, it can produce insightful network views.
DWH at Mail.ru Group — The second part covered the data warehouse architecture at Mail.ru Group: organizational structure of the DWH department, responsibilities and workflows, and the technology stack supporting one of Russia’s largest internet companies.
Why It Matters
Combining mathematical theory with practical visualization shows how data professionals can communicate complex concepts effectively. The friendship paradox is a beautiful example of how network structure creates unintuitive phenomena — and understanding these phenomena has real business value.
