Anastasiia Kornilova has over 3 years of experience in data science. She has an MS in Applied Mathematics and runs two blogs. Her interests include recommendation systems, natural language processing, and scalable data solutions. The agenda of her presentation includes defining data science, who data scientists are and what they do, and how to start a career in data science. She discusses the wide availability of data, how data science makes sense of and provides feedback on data, common data science applications, and who employs data scientists. The presentation outlines the typical data science workflow and skills required, including domain knowledge, math/statistics, programming, communication/visualization, and how these skills can be obtained. It provides examples of data science

1. Anastasiia Kornilova
1

2. WHO AM I?
• 3+ years in Data Science
• MS in Applied Mathematics
• Professional interests: recommendations systems, natural language
processing, scalable data science solutions
• Authors of two blogs: energyfirefox.blogspot.com,
datascientistdiary.blogspot.com
• Fan of online education (20+ finished MOOCs)

3. • What is Data Science and why do we need it?
• Data Scientists.Who they are and what do they
do?
• How to start?
• Practical case
AGENDA
3

4. DATA IS EVERYWHERE

5. TONNS OF DATA

6. DATA EXCHANGE

7. MAKING SENSE OF DATA

8. USERS FEEDBACK

9. DATA ISTHE NEW OIL

10. WHAT IS DATA SCIENCE?
Andrew Conway
10

11. FRAUD DETECTION

12. RECOMMENDATIONS

13. OPTIMISATION

14. SENTIMENT ANALYSIS

15. IMAGE/SPEECH/TEXT
RECOGNITION

16. FUTURE PREDICTION

17. FIND PATTERNS IS USER
BEHAVIOUR/ACTIONS

18. WHO USES DATA SCIENCE?
18

19. DATA SCIENTISTS DEMAND
19

20. WHO DATA SCIENTISTS ARE
AND WHAT DOTHEY DO?
20

21. TYPES OF DATA SCIENTISTS
A - Analysis
B - Building
Robert Chang

22. DSTYPE “A” - ANALYSIS
• making sense of data or working with it in a fairly static way.
• very similar to a statistician (and may be one)
• knows all the practical details of working with data that
aren’t taught in the statistics curriculum: data cleaning,
methods for dealing with very large data sets, visualization,
deep knowledge of a particular domain, writing well
about data

23. • share some statistical background withType A
• very strong coders and may be trained software
engineers
• mainly interested in using data “in production.”
• build models which interact with users, often serving
recommendations (products, people you may know, ads,
movies, search results).
DSTYPE “B” - BUILDING

24. WHAT DOTHEY DO?

25. Understand
Collect
Data exploration
Clean and
transform
Model
Validate
Communicating
results
Deploy
WHAT DOTHEY DO?

26. TYPICAL DATA SCIENCE
WORKFLOW
• Preparing to run a model (Gathering, cleaning,
transformation)
• Running the model
• Interpreting the results
“80% of work” - Aaron Kimball
“Other 80% of the work”
26

27. REQUIRED SKILLS
27

28. DOMAIN KNOWLEDGE AND
SOFT SKILLS
• Passionate about the business
• Curios about data
• Influence without authority
• Hacker mindset
• Problem solver
• Strategic, proactive, creative, innovative and collaborative
28

29. MATH AND STATISTICS
• Machine learning
• Statistical modelling
• Experiment design
• Supervised learning
• Unsupervised learning
• Optimisation
29

30. PROGRAMMING AND
DATABASES
• Computer science fundamentals
• Scripting language
• Statistical computing language
• Databases
• Relational algebra
• Distributed computations
30

31. COMMUNICATION AND
VISUALIZATION
• Ability to engage with senior management
• Storytelling skills
• Visual art design
• Knowledge of a vizualisation tool
• Translate data-driven insights into decisions and actions
31

32. WHERETO OBTAIN SKILLS?

33. HARD WAY
33

34. TRADITIONAL WAY
• LITS - Machine Learning
• UCU - CS Master Degree
• Data Science Degree
• Kyivstar - Big Data University
34

35. EASY WAY
35

36. MACHINE LEARNING
≠
DATA SCIENCE

38. AND NOW WHAT?
38

41. AREYOU GOOD ENOUGH?
AREYOU GOOD ENOUGH?41

42. KAGGLE STORY
42

43. Problem owners
Problem solvers
43

44. WHAT CAN YOU FIND ON
KAGGLE?
• Knowledge
• Money
• Job
• Reputation
44

45. 1. Understand
2. Collect
3. Data exploration
4. Clean and
transform
5. Model
6. Validate
7. Communicating
results
Deploy
45

46. PASSION AND PERSISTENCE

47. TIME FOR FUN
San Francisco crimes analysis

48. THANK YOU!

49. LINKS
• https://medium.com/@rchang/my-two-year-journey-as-a-data-scientist-at-twitter-f0c13298aee6#.49jdojamn
• https://blog.kissmetrics.com/how-netflix-uses-analytics/
• http://recode.net/2015/10/07/jawbone-isnt-a-hardware-company-anymore-says-ceo-hosain-rahman/
• https://jawbone.com/blog/napa-earthquake-effect-on-sleep/
• http://cs.ucu.edu.ua/
• http://lits.com.ua/course/machine-learning/
• http://bigdata.kyivstar.ua/
• https://www.kaggle.com/
• http://inversquare.github.io/moon/mooncrime.html#part-2-crimes-of-passion
• http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0059030#abstract0
• http://blog.babajob.com/wp-content/uploads/2015/08/manyhands-300x161.jpg
• http://www.criminalelement.com/images/stories/-2015-Jul-Sep/sherlock-holmes-benedict-cumberbatch.jpg