Open data science education
Jonathan Cornelissen - CEO DataCamp
8/2/2017 AncondaCON, Austin TX
Online interactive education platform for data science
~ 800,000 learners & 50M+ completed exercises
Two visions on online education
Vertically focused Generic
Languages Web dev Data Science
Scalability usually comes at the expense of personalization
Pers
onal
izat
ion
Scalability
One-on-one tutoring
Small classroom
Textbooks
Online videos
Scalability usually comes at the expense of personalization
Pers
onal
izat
ion
Scalability
One-on-one tutoring
Small classroom
Textbooks
Online videos
Do slide on other interfaces and mention NBGrader
Do slide on other interfaces and mention NBGrader
https://github.com/jupyter/nbgrader
Who wants to learn data science?
Who’s learning data science?
70%Professionals
70% of learners is younger than 35
71% of learners is male
What do they want to learn?
Confirmed by search behavior on DataCamp
Programming
Importing & Cleaning Data
Data Manipulation
Data Visualization
Probability and Statistics
Machine Learning
Reporting
Electives
R Python SQL Spark
TechnologyD
ata
Sci
ence
P
roce
ss ?
Our Python curriculum
Learning about teaching data science online
● Interactivity (and $$) correlate(s) with higher course completion rates○ Typical MOOC: <10%○ DataCamp Free Course: ~30%○ DataCamp Premium Course: 40%-75%
● R users want Python content○ ~ 40% of R students indicated they wanted Python courses
● Automating the instructor is hard -> https://github.com/datacamp/pythonwhat
Future of data science learning?
The importance of fluency
Interactive learning + repetition
=> Data Science Fluency
Practicing passive fluency
In 2017
Practicing passive fluency (2)
Moving towards practicing active fluency
Practicing active fluency
Data Science Fluency
=> Certification
2017
- Practice Mode- Mobile App- Certification- And of course, even more Python courses ;-)
Q&AJonathan Cornelissen - CEO DataCamp
[email protected]