NextGame

Insight Data Science Webapp Demo

Transcript

1. Caleb Strait
2. Which video game should I try next?
3. Which video game should I try next?
4. Which video game should I try next?
5. Which video game should I try next?
6. Have you played League of Legends? A bit, it’s not for me What about Super Mario 3D World? Yeah – what a great game! Let’s maximize per-user data by mimicking this conversation Oh! Check out the new Legend of Zelda!
8. API & web scraping Data from
9. Critic Review Scores 97 Calculating Games’ Initial Scores mean
10. Critic Review Scores 97 Calculating Games’ Initial Scores mean 92 95 …
11. Critic Review Scores 97 Calculating Games’ Initial Scores mean Display Top Game 92 95 …
12. Top Game User Feedback Adjusts Scores & Reorders Games
13. New Top Game User Feedback Adjusts Scores & Reorders Games
14. reinforcement learning similarity-weighted Algorithm:
15. Game Similarity Within-Critic Score Difference Genres
16. Normalized Average Score Difference Game Similarity 6.8 7.3 critic 1 7.8 7.3 8.6 - - - game 1 # game 2 # critic 2 critic 3 critic 4 8.1 9.2 Within-Critic Score Difference Genres
17. Normalized Average Score Difference Proportion Of Genres Overlapping Game Similarity 8.1 9.2 6.8 7.3 critic 1 7.8 7.3 8.6 - - - game 1 # game 2 # critic 2 critic 3 critic 4 game 1 # game 2 # Within-Critic Score Difference Genres
18. Normalized Average Score Difference Proportion Of Genres Overlapping Game Similarity SIM = , ) ( harmonic_mean(NASD, POGO) game 1 # game 2 # 6.8 7.3 critic 1 7.8 7.3 8.6 - - - game 1 # game 2 # critic 2 critic 3 critic 4 game 1 # game 2 # 8.1 9.2 Within-Critic Score Difference Genres
19. Cross-genre trends Overlapping fanbase
20. Caleb Strait PhD in the neural basis of economic decision-making