Diese Präsentation wurde erfolgreich gemeldet.
Wir verwenden Ihre LinkedIn Profilangaben und Informationen zu Ihren Aktivitäten, um Anzeigen zu personalisieren und Ihnen relevantere Inhalte anzuzeigen. Sie können Ihre Anzeigeneinstellungen jederzeit ändern.
Nächste SlideShare
What to Upload to SlideShare
What to Upload to SlideShare
Wird geladen in …3
×
1 von 57

Anastasia Lieva, WiMLDS Paris

0

Teilen

Herunterladen, um offline zu lesen

Anastasia Lieva, WiMLDS Paris

Ähnliche Bücher

Kostenlos mit einer 30-tägigen Testversion von Scribd

Alle anzeigen

Ähnliche Hörbücher

Kostenlos mit einer 30-tägigen Testversion von Scribd

Alle anzeigen

Anastasia Lieva, WiMLDS Paris

  1. 1. E X T R E M E D A T A - S C I E N C E @fuzzyAnastazia Anastasia Lieva
  2. 2. Once upon a time ...
  3. 3. EXTREME PROGRAMMING
  4. 4. ● Illusion of complete specifications if customer is not on site Bewitched links
  5. 5. ● Illusion of complete specifications if customer is not on site ● No one is responsible ● Lack of communication Bewitched links
  6. 6. Cost of change Time
  7. 7. ● Illusion of complete specifications if customer is not on site ● No one is responsible ● Lack of communication ● Technical debt Bewitched links
  8. 8. ● Illusion of complete specifications if customer is not on site ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  9. 9. ● Illusion of complete specifications if customer is not on site ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  10. 10. ● Illusion of complete specifications if customer is not on site ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  11. 11. EXTREME PROGRAMMING masterofproject.com
  12. 12. EXTREME PROGRAMMING masterofproject.com
  13. 13. EXTREME PROGRAMMING masterofproject.com
  14. 14. EXTREME PROGRAMMING masterofproject.com
  15. 15. EXTREME PROGRAMMING masterofproject.com
  16. 16. EXTREME PROGRAMMING masterofproject.com
  17. 17. EXTREME PROGRAMMING masterofproject.com
  18. 18. EXTREME PROGRAMMING masterofproject.com
  19. 19. Data Science comparing to Traditional Software Development?
  20. 20. Data Science comparing to Traditional Software Development? Rules Data Outputs
  21. 21. Data Science comparing to Traditional Software Development? Rules Data Outputs ML Data Rules
  22. 22. ● Illusion of detailed specifications ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  23. 23. ● Illusion of detailed specifications ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  24. 24. ● Illusion of detailed specifications ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  25. 25. ● Illusion of c̶o̶m̶p̶l̶e̶t̶e̶ ̶s̶p̶e̶c̶i̶f̶i̶c̶a̶t̶i̶o̶n̶s̶ miracles if customer is not on site ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  26. 26. ● Illusion of c̶o̶m̶p̶l̶e̶t̶e̶ ̶s̶p̶e̶c̶i̶f̶i̶c̶a̶t̶i̶o̶n̶s̶ miracles if customer is not on site ● No one is responsible ● Lack of communication ● Technical debt ● Lack of visibility Bewitched links
  27. 27. EXTREME DATA-SCIENCE
  28. 28. ● Metaphor
  29. 29. Collaborative workshops
  30. 30. ● Metaphor ● User Stories
  31. 31. The anomaly detection system Epic Build a prototype to detect the anomalies in hot water tank consumption
  32. 32. The anomaly detection system Epic Build a prototype to detect the anomalies in hot water tank consumption Story 1 As a user I can store the consumption of my hot water tank so that enable the data analysis for anomalies detection Story 2 As a expert I can label the timestamps when a hot water tank has anomaly so that the data-scientist can perform supervised learning of anomalies Story 3 As a data-scientist I can create synthetic observations targeted as anomalies so that the original imbalanced dataset not affect the machine learning performance especially regarding its sensitivity metric Story 4
  33. 33. ● Metaphor ● User Stories ● Frequent small releases
  34. 34. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity
  35. 35. The anomaly detection system Epic Build a prototype to detect the anomalies in hot water tank consumption Story 1 As a user I can store the consumption of my hot water tank so that enable the data analysis for anomalies detection Story 2 As a expert I can label the timestamps when a hot water tank has anomaly so that the data-scientist can perform supervised learning of anomalies Story 3 As a data-scientist I can create synthetic observations targeted as anomalies so that the original imbalanced dataset not affect the machine learning performance especially regarding its sensitivity metric Story 4
  36. 36. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring
  37. 37. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team
  38. 38. Workshops with end-users
  39. 39. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team ● Pair Programming
  40. 40. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team ● Pair Programming ● Coding Standards
  41. 41. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team ● Pair Programming ● Coding Standards ● Collective code ownership
  42. 42. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team ● Pair Programming ● Coding Standards ● Collective code ownership ● Test driven development
  43. 43. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team ● Pair Programming ● Coding Standards ● Collective code ownership ● Test driven development & Test EVERYTHING
  44. 44. ● Metaphor ● User Stories ● Frequent small releases ● Simplicity ● Refactoring ● Whole Team ● Pair Programming ● Coding Standards ● Collective code ownership ● Test driven development & Test EVERYTHING ● Continuous integration
  45. 45. ● Communication ● Simplicity ● Feedback ● Courage EXTREME DATA-SCIENCE
  46. 46. Thank you ! @fuzzyAnastazia Illustrations by Marina Bobyreva @marinaboby

×