Workshop: Quantitative Research with Free Software
Trujillo · viernes 21 de agosto de 2026
- Guía de una página y hoja de código R — las dos hojas que se repartieron
taller.Rydatos_ejemplo.csv— el análisis completo, para repetirlo en casa- Las diapositivas de la mañana
- El formulario, para clonarlo y cambiarle las preguntas
Guardar los dos archivos en la misma carpeta y abrir taller.R con RStudio. Viene puesta la opción C, que usa los datos de ejemplo y funciona sin internet. Todo esto se puede copiar, cambiar y volver a enseñar libremente.
A half-morning workshop for teachers at Peruvian technical institutes (institutos de educación superior tecnológica), first taught in Trujillo in August 2026 to faculty of Fe y Alegría and neighbouring public institutes.
Participants build a survey, answer it from their own phones, download the responses, and analyse and plot them — the whole cycle, once, with data they generated themselves. All the materials are in Spanish.
Two constraints shaped everything
Only free software: KoboToolbox for the survey, R and RStudio for the analysis. That is what lets the workshop keep working after the facilitator leaves.
Base R only — no install.packages(). A fresh R installation on a lab machine with weak internet cannot download ggplot2 while thirty people wait, so the script uses table(), tapply(), barplot(), hist() and boxplot() and nothing else. The constraint turned out to be a feature: everything in the script runs anywhere R runs.
Materials
| Slides | HTML · PDF |
| One-page handout | guia.pdf |
| R code sheet | hoja_de_codigo.pdf |
| The script participants run | taller.R |
| Example data (synthetic) | datos_ejemplo.csv |
| The survey, as an XLSForm | formulario_ejemplo.xlsx |
| Facilitator’s playbook | guion_del_taller.md |
The playbook is the part I would read first
Most of what makes a workshop like this succeed or fail is not in the slides. guion_del_taller.md has the minute-by-minute plan with what gets cut first when time runs short, a table of what to do when each assumption fails — no connectivity, no laptops, R not installed, the projector is VGA — and the checklist for inspecting the computer lab the day before. It is written for whoever runs this next, including me.
A few things it records that were learned the hard way: the survey has to exist before the morning starts, because the printed QR code points at it; a select_one question with appearance: autocomplete renders on a phone as an empty text box that nobody recognises as a question; and the answer file has to be opened on a phone before it is handed out, because what looks right on the projector does not always look right on a telephone.
Reuse
Everything here is CC BY-SA 4.0 — use it, change it, teach it, as long as what you pass on stays as free as what you received.
The example data is synthetic, generated by a seeded script; no real participant’s answers are published here. The institute names in the sample survey are invented.