A list of nice statistical things
When Stats4SD started its blog, I was pleased – here at last was an easy way for me to tell the world of all the important statistical things I had been thinking about. Of course, that is sort of what I had been doing since I started out in the business - a very long time ago. But this was mostly to the few people that I met and worked with; not the wider world of people engaged in statistics for sustainable development.
The scientist’s conventional way of getting a message out to many people, potentially everyone to whom it might be useful (if they took the effort to find and read it), is the scientific journal article. But writing these is such hard work, and maybe what I have to say is not original enough to be published there. Blogging has been possible for about 20 years, but there were a few steps to getting started that I was not prepared to take. Stats4SD has now made it so easy: I have no excuse!
Then I noticed that every blog piece I wrote, and many others that have been written, all say: “Look at this! So many people make this statistical mistake. Why can’t they all be statistically smart and savvy like us? Let me enlighten them and make the world a better place.”
So, inspired by an irrelevant but charming article I read this morning, I thought a bit of statistical appreciation is in order. Here are some of the ‘nice’ statistical things I have noticed recently:
- Award-winning stats communication, such as that found every month in Significance.
- Graphs are a central part of communicating data and there are so many superb examples It gets easier and easier for anyone to make these – you no longer have to be that rare combination of statistician, artist and software engineer.
- In R, we have statistics software of extraordinary power available free and open source; for anyone to use or build on. It’s growing all the time, and you find many of the recent developments in statistics methods available there. And it’s all built by a community of dedicated people who make their efforts available to the world.
- To help anyone using R, there is a vast array of material available online – guides, lecture notes, presentations, videos, examples… and numerous groups that offer help and suggestions to anyone that asks (politely).
- Data data data… The very rapid growth in the volume of data about almost anything is not new news. Along with it, there is a boom in adopting open data standards, which means there is more available to people who need it. No longer is there the excuse for those teaching and learning statistics to have no data to work with.
- It’s great if the data you need exists, but of course new measurement is always going to be needed to meet new needs. We are seeing the development of methods and approaches for measuring and collecting data in areas that have always seemed difficult. For example, I work with groups using the Sustainable Intensification Assessment Framework that sets out a practical approach for integrated assessments of smallholder farms.
- Along with free open source software for statistics, we have similar initiatives, equally effective, in other areas of data use. In my work, I use spatial information and make use of QGIS; ‘the R of geographical information’.
‘All the nice statistical things’ – image courtesy of Nicolas Greliche @Stats4SD
I could go on with the list of nice statistical things in the bigger picture of statistics and sustainable development, but wanted to also list a few of the more personal nice things that I have noticed recently. Without going into detail, I include here:
- Individual students who are becoming skilled in making sense of messy data - and understanding what can and cannot be inferred. More importantly; being able to explain what they have and have not managed to find out.
- Farmers understanding important ideas about experimental design so they can take part in designing their own useful experiments.
- A student I worked with winning an award for their research.
- A project I work with getting some very positive reviews from its funder.
- A colleague who spotted an intriguing question emerging from their data analysis - asking it - and finding that I could answer it.
I could go on, but I hope the point is clear: we, the data people, aren’t only engaged in finding fault with others’ work.
Does anyone else have nice statistical things they would like to highlight? I’d be keen to hear ?