Teaching R as a GIS: problems, solutions and lessons learned

Slides presented by myself and Rachel Oldroyd at the FOSS4G conference in Como, Italy. See http://europe.foss4g.org/2015/ for more details of the conference.

Transcript

1. Teaching R as a GIS: problems, solutions and lessons learned Robin Lovelace and Rachel Oldroyd University of Leeds
2. • Statistical programming language and data environment • Based on the statistical language S • Can rival proprietary GIS's due to the emergence of advanced spatial packages Background
3. Advantages • Reproducibility • Batch Processing • Extensible environment – packages / libraries • Fast • Advanced Graphics
5. Slow uptake amongst social scientists – command line interface (CLI) – the popularity of traditional GIS (QGIS, ArcMap, MapInfo) – Pre-conceptions, steep learning curve – Failure to see usefulness – Absence on UG /PG curriculum Barriers to Learning
6. Aims and Objectives • Raise awareness of R's spatial capabilities • Teach new methods to social scientists – TALISMAN project (2010-2015) – Consumer Data Research Centre (2015 - ) • Promote the use of Open Source Software
7. The Teaching Activity • 1 day practical workshop • Variety of audiences • Over an 18 month period • Formal and informal Likert feedback collected • 'Experiential learning'
8. Difficulties • Variety of participant prior knowledge – 3 'types of participant' • Material Covered – Originally 2 tutorials – 'I was acting more like a typewriter' – 'I was tying code without understanding what I was doing'
9. Difficulties • Technical Difficulties – Permissions • Finishing early
10. Solutions I • Quality not quantity of materials – Focus on 1 tutorial • Extension tasks – 'The course strikes a good balance...for an R novice and interesting for more experienced R users' – 'The pace was just enough and there were enough extra bits of information to keep me interested'
12. Solutions II • Supportive learning environment – High number of demonstrators – Interactive • Use of IDE - RStudio – Autocompletion – Help – Plots – Data Environment
13. Solutions II • Use of traditional GIS – QGIS – For those not familiar with geographical datatypes – Visualisation of attribute and geometry data
14. Results
15. Lessons Learned • Start at the beginning • Evolving, dynamic materials • Communication = vital
16. Lessons Learned • Dissemination – CRAN, Twitter, Conferences • Clearly define target audience – Not suitable for some participants – Clearly identify prerequisites beforehand
17. Reference Lovelace, Robin, and James Cheshire. "Introduction to visualising spatial data in R." (2014) http://robinlovelace.net/r/2014/01/30/spatial- data-with-R-tutorial.html