This is my talk from the PyDataLondon conference in May 2016. I outline some time management techniques and useful learning resources for those interested in transitioning into data science.
4. …WHY WOULD IT EVEN
WORK AS AN APPROACH? -
SOME HANDWAVY EXAMPLES
5. HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
WHY DOES OUR SOCIETY SUPPORT HIGHER EDUCATION?
▸ Social pressure insufficient to explain equilibria
▸ Historical changes
▸ Robin Hanson: it’s not about social pressure, it’s about
prestige/status signalling
6. PRESTIGE IS EXPLOITABLE
▸ Public and private tuition costs are skyrocketing (USA market)
Source: http://www.theatlantic.com/education/archive/2015/05/the-real-cost-of-college/393086/
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
7. PRESTIGE IS EXPLOITABLE
▸ …without providing much real value
Best predictors of employee success according to meta-analysis performed by
Schmidt and Hunter:
▸ 1. Work sample tests (.54)
▸ 2. GMA tests aka IQ (.51)
▸ 3. Employment interviews -- structured (.51)
▸ … 14. Job experience --years (.18)
▸ … 16. Years of education (.10)
Source: Schmidt, F.L. & Hunter, J.E. (1998) The validity and utility of selection methods in personnel psychology: Practical and
theoretical implications of 85 years of research findings,” Psychological Bulletin, 124, 262–274.
Article: http://bobsutton.typepad.com/my_weblog/2009/10/selecting-talent-the-upshot-from-85-years-of-research.html
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
9. WHAT DID IT MEAN FOR ME?
▸ Cost of a Master’s degree for a non-EU student: £28k
▸ Starting date: October 2016 (if accepted!)
▸ Funding prospects for PhD: ahem… slim.
▸ Improved career prospects?
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
10. SO, THINGS NOT TO DO IF YOU WANT A RADICAL CAREER CHANGE
▸ Hope that degree is a ticket to a rosy future
▸ Expect to do it all on your own
▸ Not allow for a margin of error
▸ Burn your bridges
▸ Hope that employer can read your mind
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
11. WHAT DID I DO IN THE END?
▸ Set a realistic timeframe
▸ Don’t trust yourself with sticking to deadlines
▸ Make a study plan
▸ Prepare for uncertainty
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
12. TIME DISCOUNTING AND WILLPOWER
▸ Willpower is a limited resource
▸ External reinforcement and social cost of default
▸ Limit your choices
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
13. TIME MANAGEMENT TECHNIQUES
▸ Pomodoro
▸ Spaced repetition
▸ Value of sleep
▸ Every little counts. Even 5 minutes!
▸ Kill your perfectionism
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
16. NETWORKING FOR NERDS
▸ Understand your in-group
▸ Signal belonging
▸ Positive reinforcement
▸ This is still work!
▸ Don’t expect immediate results
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
17. NETWORKING FOR NERDS
▸ Understand your in-group
▸ Signal belonging
▸ Positive reinforcement
▸ This is still work!
▸ Don’t expect immediate results
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
18. RESOURCES FOR DATA SCIENCE TRANSITION
▸ Meetups! PyData, LMLP, Data Science journal Reading club, etc. Or
start one yourself!
▸ Kaggle and hackathons
▸ Coursera
▸ Know your stats. “Elements of Statistical Learning”, “An Introduction to
Statistical Learning with Applications in R”, Jaynes
▸ Pester people!
▸ Read the white papers. And the blogs. DeepMind, Karpathy, Google
it!
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
19. DON’T GET STARTED
▸ Complex projects require extended planning
▸ You need the right questions
HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
21. HOW TO BECOME A DATA SCIENTIST IN 6 MONTHS
SOME LEARNING RESOURCES
‣ Andrew Ng’s Coursera: https://www.coursera.org/learn/machine-learning
‣ TensorFlow course on Udacity: https://www.udacity.com/course/deep-learning--ud730
‣ Stanford course on Convolutional NNs: http://cs231n.stanford.edu/syllabus.html
‣ Kaggle. In case you haven’t heard :) https://www.kaggle.com
‣ Fellowships: S2DS ( http://www.s2ds.org ), ASI ( http://www.theasi.co )
‣ Meetups: PyData (duh), London Data Science (#ODSC), London Kaggle Meetup, London Machine Learning Practice
Meetup (LMLP), London Machine Learning Meetup (different meetup from LMLP but both hosted at CodeNode!),
London Data Science Journal Club, events at Google Campus, and many many more.
‣ Books: Probabilistic Theory of Pattern Recognition ( http://www.szit.bme.hu/%7Egyorfi/pbook.pdf ), Elements of
Statistical Learning ( http://statweb.stanford.edu/~tibs/ElemStatLearn/ ), Probability Theory by ET Jaynes ( http://
bayes.wustl.edu/etj/prob/book.pdf - note this is not the whole text)
‣ A more comprehensive list of MOOCs in a Google Doc (second tab): https://docs.google.com/spreadsheets/d/
1J0sRe-84ynJI1LF8XHESAgnMNNc1c7lmf-HGsCrcbm8/edit?usp=sharing
‣ Get involved: Machine Learning For Social Good Facebook group https://www.facebook.com/Machine-learning-for-
good-259545294213595/