training / result reports
Result reports that write themselves.
New data? One click – and every figure, table and chart in the report is up to date and traceable. That is good scientific practice – and it saves hours every time the report comes round again.
Don’t worry: this is not a programming course. You can even start with no code at all, using Excel pivot tools.
A miniature R notebook to play with: click the dashed values in the code, run the chunk – and the chart below obeys. That is exactly how working in R feels.
focus
The three strands of the course
The aim: to put results into a report without copy and paste, in a way that makes it traceable at any time where a figure or a chart came from. If data collection is still running, you write the report anyway and bring it up to date with one click once the data change.
Preparing the data
Whether Excel or R Markdown – both need data in a workable shape. We go through the main dos and don’ts: one row per case, clean column names, consistent codes, and what LimeSurvey exports contribute.
Reports with Excel pivot tools
For descriptive statistics: automated result reports built on pivot tables – carried through to Word or PowerPoint if you wish. Little known: charts can be inserted so that they update in PowerPoint as soon as something changes in Excel. That saves a great deal of work on recurring reports – but it has pitfalls, because a slide can then also change by accident.
Reports with R Markdown
An introduction to R Markdown: reports whose figures come straight from the data – through to complete theses, presentation slides or websites produced from RStudio (see my freely available project “Thesis Writing in RStudio”).
Formats: half a day or a full day, online or in house – up to 6 participants in the standard format, larger groups quoted individually.
live_demo / why_r_markdown
Why R Markdown? See for yourself.
On the left, an R Markdown script for a course feedback report – the green passages are R code in the middle of the text. Click “Knit” and the finished report appears on the right, every figure calculated directly from the data. And now the real trick: choose the next wave of data collection and knit again – the number of participants, the means, the table, the chart and even the wording adjust by themselves. No copy & paste, no forgotten figure.
This is exactly the principle – write the script once, have the report for good – that I build up with your team in training courses, for your own surveys, including the ones that come round every year.
ai_and_code
Have AI write the code – and be able to judge it.
The AI writes your R code. “Change the bar colour to blue” is enough, and you get an amended line back. That is precisely where the difference from Excel lies: in R the change is a piece of text that you can read, check and undo. In Excel it is a sequence of clicks – an AI can describe them, but nobody can later see from the file what was changed.
Even so: for a one-off small analysis Excel is quicker, and an AI guides you through click paths perfectly well. R pulls ahead as soon as something has to be repeated, checked, or followed by somebody else.
And your data stay where they are: the analysis runs locally on your own machine – the AI sees the script, not your data. So you need neither be afraid of code nor worry about data protection when you get help with an analysis script: “I would like to produce a histogram with blue bars in R” is instruction enough – any AI chat manages that, including the ones behind no paywall.
There is a second benefit that many people overlook: AI models answer in Markdown. Headings, lists, tables, code blocks – anyone who understands that format can reuse the answers cleanly instead of laboriously taking them apart. So what you learn about Markdown here helps you outside R as well.
quote
From data set to report
Describe your data and your goal in a few lines – I will suggest a training or consulting format that fits.
Enquire without obligation