By Simon Hettrick

Deputy Director, Software Sustainability Institute

What! We have to worry about software now?

6 November 2018 - ESRC Festival     @sjh5000      ORCID: 0000-0002-6809-5195

I'm not a social scientist

These slides: bit.ly/2PF6FS8

My skills

Laser

physics

These slides: bit.ly/2PF6FS8

"How many researchers rely on software?"

These slides: bit.ly/2PF6FS8

These slides: bit.ly/2PF6FS8

My skills

Laser

physics

Social

science

These slides: bit.ly/2PF6FS8

"How many researchers rely on software?"

These slides: bit.ly/2PF6FS8

Use

software

Fundamental to

results

Develop own code

69%

92%

56%

These slides: bit.ly/2PF6FS8

Percentage of cross-council funding into software-reliant research

These slides: bit.ly/2PF6FS8

Software in publications

65%

31 UK institutions ~ 600k papers

These slides: bit.ly/2PF6FS8

Software is rapidly becoming a fundamental research tool

These slides: bit.ly/2PF6FS8

All good so far...

These slides: bit.ly/2PF6FS8

of researchers who develop

code have no training

24%

of group leaders looking to commercialise code would not feel comfortable sharing their code

50%

REF 2014 publication submissions

>15,000

REF 2014 software submissions

58

These slides: bit.ly/2PF6FS8

You can't have reliable results without reliable software

 

You can't reproduce results without access to software

These slides: bit.ly/2PF6FS8

Researchers can't afford to care while research incentives are against them

These slides: bit.ly/2PF6FS8

What's the situation in social science?

These slides: bit.ly/2PF6FS8

Percentage of ESRC funding into

software-reliant research

These slides: bit.ly/2PF6FS8

Why?

  • Data deluge

  • More use of data not designed for social science

  • Lots of data cleaning and data alignment problems

  • More software techniques, machine-learning, automated data collection, etc

  • Lots of Excel, Stata and SPSS, less OpenRefine, R and Python

  • Training in the traditional tools, but little more advanced training​

These slides: bit.ly/2PF6FS8

What do we

need to do?

These slides: bit.ly/2PF6FS8

Incentivise good software practices

These slides: bit.ly/2PF6FS8

Incentives

  • Submit software to REF2021!

  • Better assessment of grants that include software development

  • Connecting research outputs to grants that include software development

  • Software management plans

These slides: bit.ly/2PF6FS8

Make training available

These slides: bit.ly/2PF6FS8

Growth of carpentry workshops in the UK

Data Carpentry for social science: http://bit.ly/2PeMOda

These slides: bit.ly/2PF6FS8

Make skills

available

These slides: bit.ly/2PF6FS8

RSE Groups at

20 organisations

bit.ly/RSEGroupsInUK

These slides: bit.ly/2PF6FS8

www.rse.ac.uk

Software use is only going to increase

 

And this is a good thing!

 

Expect more pressure to adapt research practices

These slides: bit.ly/2PF6FS8

Studies

  • Software survey:
    https://www.software.ac.uk/blog/2017-09-06-journey-reproducibility-excel-pandas
     
  • Grants data analysis:
    https://github.com/softwaresaved/software_in_grants_GTR
     
  • Publication analysis coming soon

 

Contact

s.hettrick@software.ac.uk

@sjh5000

ORCID: 0000-0002-6809-5195

Licence

 © Simon Hettrick 

These slides are licensed under a Creative Commons Attribution 4.0 International 

https://creativecommons.org/licenses/by/4.0/

Thanks

Peter Smyth and Aleksandra Nenadic

These slides: bit.ly/2PF6FS8