Play Kaggle with R, Facebook V: Predicting Check Ins
2 likes673 views
The document discusses the use of R for participating in Kaggle competitions, particularly focusing on predicting Facebook check-ins. Mia Chang presents the R-Ladies Taipei group, introduces data analysis through exploratory data analysis (EDA), Azure Jupyter Notebooks, and emphasizes sharing resources among the community. Key takeaways include utilizing EDA for feature extraction and accessing various Kaggle resources.
Play Kaggle with R, Facebook V: Predicting Check Ins
- 1. 用 R 玩 Kaggle – 臉書打卡點預測 Play kaggle with R, Facebook V: Predicting Check Ins @Mia (R-Ladies) library(dplyr) r-ladies_global %>% filter(from = 'Taipei', travel_to = 'Lisbon')
- 2. The Agenda First Second Third Hey, Kaggle # with R-Ladies # with Masters Play R # Warm Up # EDA, Shiny Apps # Azure Jupyter Notebook Brief Intro # About R-Ladies # About me 2 Last Q&A # Recap # Resource Sharing
- 3. 1. Brief Intro Intro about R-Ladies and me! https://rladiestaipei.github.io/R-Ladies-Taipei/ 3
- 4. R-Ladies Taipei 4 Berl in Taiwa n 4 ● Founded in 2014 ● every month in Taipei, ● Our history: goo.gl/HbHNeP
- 5. Hello! I am Mia Chang (張懷文). ▪ Data Scientist, Lecturer ▪ Member of R-Ladies Taipei ▪ Co-founder of Azure Taiwan Community ▪ Microsoft Most Valuable Professionals (MVP) 2017 5
- 6. 2. Hey, Kaggle With R-Ladies and other masters! 6
- 7. 7
- 8. 8
- 9. 9
- 10. 3. Play R # Warm Up - 問題背景，問題定義 # Azure Jupiter Notebook # use Jupiter access Data # 結論 10
- 12. Warm Up - 關於這個問題背景，問題定義 12 With 8.6 million test records there are about a trillion (10^12) place-observation combinations.
- 13. Warm Up - 關於這個問題背景，問題定義 Schema row_id x y accuracy time place_id 13 EDA
- 14. “ Warm Up - 關於這個問題背景，問題定義 Three weeks into the eight-week competition, I climbed to the top of the public leaderboard with about 50 features 1. the summary data such as the number of historical check ins. 2. historical density of a place candidate, one year prior to the observation. 3.All features are rescaled if needed in order to result in similar interpretations for the train and test features. 14
- 16. EDA - 發現大家透過GPS, Wi-Fi or cellular 16 c
- 17. EDA - 尖峰打卡日 17
- 18. 玩資料 - Shinny App 18
- 19. 玩資料 - Github 19
- 20. 玩資料 - Azure Jupyter Notebook 20
- 21. 玩資料 - Azure Jupyter Notebook 21
- 22. # 演算法及結論 ＃Rcpp ＃It was expected that it would be clearly correlated with the variation in x and y but the pattern is not as obvious. Halfway through the competition I cracked the code ... 22
- 23. 4. Q&A # Recap # Action Item 23
- 24. Recap First Second Third Hey, Kaggle # with R-Ladies # with Masters Play R # Warm Up # EDA, Shiny Apps # Azure Jupyter Notebook Brief Intro # About R-Ladies # About me 24 Last Q&A # Recap # Resource Sharing
- 25. Action Item First Second Third Hi, Kaggle Play R Get your partners Visit R-Ladies R-Basic too! 25 Then ...
- 26. Thanks for your listening! 26 Look forward to your visit to R-Ladies Taipei! Also Azure Taiwan!
- 27. Bye! I am Mia Chang (張懷文) ▪ mia5419@gmail.com ▪ facebook.com/mia5419 27
- 28. 28 Take Away & Reference 1.Use EDA to help you find more feature. 2.Go to Kaggle website to get more resource to help you: forum, kernels 3.No matter you are learning R or you are going to traveling to visit other R-Ladies, call us for more resources :) 1. R-Ladies Meetup Page 2. R-Ladies Facebook Group 3. Blog Post by Tom Van de Wiele - Detail about implementation 4. Github Repository 5. Shiny App by Tom Van de Wiele - EDA that you can learn more 6. Kaggle Event Page 7. Microsoft Azure Notebooks