how to build a matlab community melbourne · data science jobs (% of total jobs) 0.1 0.01 4....

21
How to build a MATLAB community Kerry Halupka | @EngKerry | [email protected] Research Platform Services | The University of Melbourne

Upload: others

Post on 17-Jul-2020

7 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

How to build a MATLAB communityKerry Halupka | @EngKerry | [email protected]

Research Platform Services | The University of Melbourne

Page 2: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Problem

The amount of stored data grew by 23% a year between 1986 and 2007

Hilbert & Lopez, Science 332.6025 (2011)

• Modern research relies on digital tools

2

Page 3: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Problem • More devices communicating = more languages for researchers to learn

3

Page 4: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Problem • High staff / student turnover in the lab

Data sc

ience jobs (%

 of total jobs)

0.1

0.01

4

Page 5: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Problem

• Learning to program is

slow

• Incentives to teach are low

• Burden falls to individuals

5

Page 6: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Problem

• Learning to program is

slow

• Incentives to teach are low

• Burden falls to individuals

Unhappy, inefficient researchers

6

Page 7: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Research Bazaar

• A community based around learning to code

7

Page 8: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Community

Bazaar – open network Cathedral – hierarchy

8

Page 9: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Community

Bazaar – open network

9

Page 10: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Community

Open Engaging Fun Long-term

10

Page 11: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Openess

• Courses are designed and run

by researchers, for researchers

11

Page 12: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Openess

• Course notes are published

online

• Fits well with the online

MATLAB community

12

Page 13: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

13

Page 14: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Engagement

• Run weekly eventsHacky Hour

PhTea

14

Page 15: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Engagement• Actively target diverse groups

• Seek new volunteers after

every course

15

Page 16: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Make it fun

• Challenge based learning

• Games / Codathon

• Mascots everywhere

16

Page 17: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Make it fun

• Cody coursework

• Cody challenges

17

Page 18: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Long-term

• Promote the community

• Have a community space

• Social media groups

CoLab

18

Page 19: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Results• 10 MATLAB Courses

• 300+ participants

• 30 volunteers

• Extension Courses – Statistics & Machine Learning, MDCS, Image Processing

• Over 50% women involvement

• Seeing the community grow internationally

19

Page 20: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

ResBaz Digital Research Conference

20

Page 21: How to Build a MATLAB Community MELBOURNE · Data science jobs (% of total jobs) 0.1 0.01 4. Problem • Learning to program is slow • Incentives to teach are low • Burden falls

Thanks for listening!

Acknowledgements:

David FlandersPhilippa Karoly& the ResBaz team

@ResPlat

21