Presentation is loading. Please wait.

Presentation is loading. Please wait.

Showing your working: a how to guide to reproducible research

Similar presentations


Presentation on theme: "Showing your working: a how to guide to reproducible research"— Presentation transcript:

1 Showing your working: a how to guide to reproducible research

2 Research fellow at the Alan Turing Institute for Data Science
Research associate in the Brain Mapping Unit, Department of Psychiatry, University of Cambridge 2016/17 Mozilla Fellow for Science

3 Reproducible vs Replicable

4 Data Same Different Code

5 Data Same Different Code Reproducible

6 Data Same Different Code Reproducible Replicable

7 Data Same Different Code Reproducible Replicable Robust

8 Data Same Different Code Reproducible Replicable Robust Generalisable

9 Barriers to reproducible research

10 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

11 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

12 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

13 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

14 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

15 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

16 Barriers to reproducible research
Is not considered for promotion Requires additional skills Plead the 5th Barriers to reproducible research Support additional users Takes time Held to higher standards than others Publication bias towards novel findings

17 Start small

18 Protocols.io

19

20

21 Coding your analyses

22 Comments are your friend!

23

24 Aim for 40% comments in your code

25 Share your comments with the original author

26 (which is almost always going to be YOU!)

27 https://twitter.com/kcranstn/status/370914072511791104

28 What if I can’t share my code until I’m published?

29

30 Python vs R (vs Matlab vs STATA etc…)

31

32 Shiny gallery

33 Anaconda

34 Click each example to see source code
Matplotlib gallery Click each example to see source code Matplotlib & Seaborn

35

36 Jupyter Notebook

37

38 Asking for help: StackOverflow & MWEs

39 https://mrcolley.com/2014/07/01/
talk-to-the-duck-debugging-and-resilience

40 Version Control

41

42

43 Some jargon busting Clone Repository Pull request Branch
Version control Merge Issues Release Commit Fork Markdown

44 Submit your first pull request!
Inspired by:

45 Submit your first pull request!
Inspired by:

46 Community https://www.mozillascience.org http://software-carpentry.org

47 Thank you! KirstieJane/ReproducibleResearch @kirstie_j
doi: /m9.figshare


Download ppt "Showing your working: a how to guide to reproducible research"

Similar presentations


Ads by Google