ビックデータとPythonではじめる野球の統計分析 #pyconjp
4 likes10,041 views
This document summarizes a presentation given at PyCon JP 2016 about analyzing baseball data with Python. The presentation introduced the speaker, Shinichi Nakagawa, and discussed using the MLBAM dataset and Python libraries like pandas and matplotlib to analyze pitching data. Specific examples analyzed the pitching of Yu Darvish before and after Tommy John surgery, compared Ichiro Suzuki and Joey Votto's batting, and looked at pitch location data to study the strike zone. The presentation emphasized the usefulness of Python for accessing and analyzing sports data.
More Related Content
Similar to ビックデータとPythonではじめる野球の統計分析 #pyconjp (6)
More from Shinichi Nakagawa (20)
Recently uploaded (14)
ビックデータとPythonではじめる野球の統計分析 #pyconjp
- 1. Python Analyzing Baseball Data With Python PyCon JP 2016 Talk Session 2016/9/22 Shinichi Nakagawa, Ai Makabi
- 2. Starting Member • Who am I ? • Hack→! Hack!! • (MLBAM) Python & (pitchpx) • Jupyter lab, pandas, matplotlib • LT • • VS ※ • @amacbee •
- 3. Who am I? • Shinichi Nakagawa(@shinyorke) • Python ※ • visasQ( ) Web Engineer • Pythonista, Agile & Scrum & Kanban coach, Baseball Analyst • / • • ( ) • 8-7 9 .
- 4. visasQ? • . • 1 2 . • Django . • https://service.visasq.com • CTO LT (ry
- 5. hack? • ( ) Python OSS . • ” ” . Web ( • . → ! • (@shinyorke) /MLB .
- 6. Hack! Python PyCon JP 2015 http://www.slideshare.net/shinyorke/hackpython-pyconjp
- 7. Hack! Python PyCon JP 2015 http://www.slideshare.net/shinyorke/hackpython-pyconjp
- 8. Hack! Python PyCon JP 2015 http://www.slideshare.net/shinyorke/hackpython-pyconjp
- 9. Hack!→ Hack!! • etc… • • ( ) ( , etc…) • … •
- 10. MLB at BAT MLB • MLB • PC TV • MLB.TV •
- 11. MLBAM Dataset(Pitch f/x) • MLB Advanced Media(MLB ) XML . • PITCH f/x Statscast Sensor ( / ) . • ( )100 . • Copyright , OK. • Github . • http://gd2.mlb.com/components
- 12. 1 (2016/4/6 VS ) http://gd2.mlb.com/components/game/mlb/year_2016/month_04/day_06/ gid_2016_04_06_lanmlb_sdnmlb_1/inning/inning_1.xml ※ pitcher=“628317"
- 13. MLBAM Dataset( / ) • etc… • • ( ) • (Relation) • (≒ ).
- 15. Analyzing Baseball Data with R • MLBAM Retrosheet(※ Hack ) • R(R Studio, pitchRx) • https://www.amazon.co.jp/dp/B00GBC36S4
- 16. PyCon Python
- 17. pitchpx - Getting MLB dataset • MLBAM XML CSV . • pitchRx(R) (@shinyorke) . • Python 3.3.x ( :Legacy Python!!!). • BeautifulSoup( ),click( ) . • PyPI . • Jupyter, pandas .
- 18. $ # Python 3.3 ( Python 3.4 ) $ pip install pitchpx $ # 2015/8/1-8/12 $ pitchpx -s 20150801 -e 20150812 -o .
- 19. Python • R Python . • (Python) • Web API Python . • Hack
- 20. • Jupyter + pandas + matplotlib , PyData . • Jupyter lab . • INPUT/OUTPUT, pandas. • matplotlib,seaborn . • Qiita, .
- 21. (Python ) …
- 23. LT • Yu Darvish (@shinyorke) • 2016 (@amacbee) ※ ( ) • VS (@shinyorke)
- 25. Yu Darvish( ) • • • (2015) TJ( )
- 26. Ichiro Suzuki( ) • & . • ( ) . • .
- 27. Joey Votto( ) • (2007-2015) . • . • 51 PyCon ( ).
- 28. Yu Darvish
- 29. Yu Darvish(2013, 2016) • TJ( ) ( etc…) • TJ (2013) (SL) (FC) • TJ (2016) 4 (FF) 2 (FT) SL FC • !? 2013(TJ ) 2016(TJ ) ※ 'CH': 'Change-up', 'CU': 'Curveball', 'EP': 'Ephuus', 'FA': 'Fastball', 'FC': 'Cut Fastball', 'FF': 'four-seam Fastball', 'FO': 'Forkball', 'FS': 'Split-finger Fastball', 'FT': 'two-seam Fastball', 'KC': 'Knuckle Curve', 'KN': 'Knuckleball', 'SC': 'Screwball', 'SI': 'Sinker', 'SL': 'Slider', 'UN': 'Unknown'
- 30. 8 • http://www.sponichi.co.jp/baseball/news/2016/08/03/kiji/K20160803013085860.html • ( ) ( ) • ?
- 31. 5 7 8 • ( DL ) • 5-7 8 • 4 2016/5-7 2016/8 ※ 'CH': 'Change-up', 'CU': 'Curveball', 'EP': 'Ephuus', 'FA': 'Fastball', 'FC': 'Cut Fastball', 'FF': 'four-seam Fastball', 'FO': 'Forkball', 'FS': 'Split-finger Fastball', 'FT': 'two-seam Fastball', 'KC': 'Knuckle Curve', 'KN': 'Knuckleball', 'SC': 'Screwball', 'SI': 'Sinker', 'SL': 'Slider', 'UN': 'Unknown'
- 32. Four seam • TJ 4 !? • 2013 (called strike) • 2016 (swinging strike) • 2013(TJ ) 2016(TJ )
- 33. Ichiro VS Votto
- 35. ( )
- 36. VS (2016 ) • • (FF) • (FT) ( ) • (42) (33) !? Ichiro Suzuki Joey Votto ※ 'CH': 'Change-up', 'CU': 'Curveball', 'EP': 'Ephuus', 'FA': 'Fastball', 'FC': 'Cut Fastball', 'FF': 'four-seam Fastball', 'FO': 'Forkball', 'FS': 'Split-finger Fastball', 'FT': 'two-seam Fastball', 'KC': 'Knuckle Curve', 'KN': 'Knuckleball', 'SC': 'Screwball', 'SI': 'Sinker', 'SL': 'Slider', 'UN': 'Unknown'
- 37. ? (20 ) ??
- 38. • ( 165cm) 2 • … (SL). • ?!?!?!? • !?( ) Jose Altuve(26) Kris Bryant(24) ※ 'CH': 'Change-up', 'CU': 'Curveball', 'EP': 'Ephuus', 'FA': 'Fastball', 'FC': 'Cut Fastball', 'FF': 'four-seam Fastball', 'FO': 'Forkball', 'FS': 'Split-finger Fastball', 'FT': 'two-seam Fastball', 'KC': 'Knuckle Curve', 'KN': 'Knuckleball', 'SC': 'Screwball', 'SI': 'Sinker', 'SL': 'Slider', 'UN': 'Unknown'
- 39. × ×
- 40. VS Part2(Hit Zone) • X (start_speed, ), Y , (pa_event_cd) • pa_event_cd (20), (21), (22), (23) ※Retrosheet • pitch_type orz ※ 'CH': 'Change-up', 'CU': 'Curveball', 'EP': 'Ephuus', 'FA': 'Fastball', 'FC': 'Cut Fastball', 'FF': 'four-seam Fastball', 'FO': 'Forkball', 'FS': 'Split-finger Fastball', 'FT': 'two-seam Fastball', 'KC': 'Knuckle Curve', 'KN': 'Knuckleball', 'SC': 'Screwball', 'SI': 'Sinker', 'SL': 'Slider', 'UN': 'Unknown' Ichiro Suzuki Joey Votto
- 41. VS Part2(Hit Zone) • X (start_speed, ), Y , (pa_event_cd) • pa_event_cd (20), (21), (22), (23) ※Retrosheet • pitch_type orz ※ 'CH': 'Change-up', 'CU': 'Curveball', 'EP': 'Ephuus', 'FA': 'Fastball', 'FC': 'Cut Fastball', 'FF': 'four-seam Fastball', 'FO': 'Forkball', 'FS': 'Split-finger Fastball', 'FT': 'two-seam Fastball', 'KC': 'Knuckle Curve', 'KN': 'Knuckleball', 'SC': 'Screwball', 'SI': 'Sinker', 'SL': 'Slider', 'UN': 'Unknown' Ichiro Suzuki Joey Votto
- 42. • 4 • ( ) … • , (& ISSUE )
- 45. “”” , . , , . “”” -
- 47. • Kawasaki.rb #kwskrb http://kawasakirb.github.io/ PyCon JP 2014 2016, • BPStudy #bpstudy http://bpstudy.connpass.com/ BPStudy Hack . •
- 48. & PyCon JP 2016 ! Shinichi Nakagawa(Twitter/Facebook/visasQ:@shinyorke)