I am at a stage where I am beginning to write up my results and analyse my data. I understand my own methodology etc., but I am struggling with creating plots and graphs that both look good and get the information across in a clear way. My supervisors are excellent sources of advice but they prefer to leave stylistic choices up to me and comment only on the content.

Does anyone have any advice or recommendations of good websites, books etc., that give advice on creating plots and graphs? I see it is a weakness not just in my own work but in other students in my department, but I'm not sure where to start in looking for advice.

To clarify, I don't just mean labelling axes, annotating with text etc. My question is quite broad and covers everything from knowing how to choose the right graph, to making it look visually appealing, and to ordering those graphs in the right order so that they continue to tell a clear and logical "story" about the results. There is so much advice on writing a literature review or choosing the right statistical test, that it is surprising how difficult I am finding it to find clear and useful advice on visualising data for a PhD.

For your actual thesis, I'd follow conventions in your field - just look through journals to see how data is presenting. Academics are stuffy and hate change :p so its a safe bet to stick with what your examiners will know. Anything too 'fancy' or modern could mean amendments - that's my thinking anyway.

For other audiences, I just think about having to explain it to my mum! So I draw whatever diagram makes the most sense in that scenario - sometimes i'll add more complicated info to stop "yeah, but what were your beta values" kinds of questions

======= Date Modified 03 Jul 2012 17:35:24 =======

Thank you, Sneaks.

Actually, my situation is slightly complicated by the fact that I am in a multidisciplinary science, spanning three different fields. All three fields have very different standards of what is justifiable science and what isn't. I keep trying to tailor my analysis to start at the simplest view and lead rapidly up to the more complicated work to avoid alienating anyone. It can be very difficult trying to balance the conventions of three different fields, particularly as many other theses on the same topic generally lean more towards one of the parent sciences or another.

Here is a question: how can you come up with a hundred plots, all in greyscale, and make each one look interesting enough to captivate the audience's attention? At least a quarter of my plots are showing changes in numerous variables over time. These kinds of plots do frequently appear in journals, but I have so many of them for my thesis that it is difficult to see how any examiner is going to read through them all without getting frustrated or bored. I could start to summarise them in different ways, but then, as you point out, I would then be using methods not common in my field.

Do you have an external examiner lined up, or any inkling which field they would come from? - if so I'd use that to guide you which discipline to follow more closely.

Do you need all the plots - my subject isn't particularly mathematical (apart from social science type stats), but IMO less is more.

Also have you thought about more multimedia type ways of presenting the data e.g. youtube vid of slide changes or something? I'm sure these things are becoming more acceptable.

Hi Sneaks,

Thanks, I am hoping to use multimedia for a portion of my thesis which would be impossible to fit into an appendix. It works much better as a video but I haven't got approval from my school yet. For the other work, unfortunately it does need to be done because each one provides the framework for the more complex analysis. I need the basic results per experiment in order to build the remaining chapters. 100 plots was an exaggeration, there are actually only about 40, and I can scale them down to fit at least 4 per page (one main large and three small), which still leaves about 10 near-identical pages of graphs, all in greyscale, which is going to be very dull.

I haven't chosen an external examiner yet because my supervisor and I want to see where my results go. My examiners don't need to be notified until January so I have time. I have chosen one in my mind, and keep trying to see my work from his viewpoint, but I don't want to cause problems for myself if it ends up being a different examiner.

In relation to presenting your data publicly, what about prezi as an alternative to powerpoint? Registration is free for those in an educational setting. I know it's the 'thing' of the moment but prezi presentations look good and are quite interactive. If not prezi or similar I would try to upskill on powerpoint. Most people use it in a quite boring way but there is a lot that can be done with it to make your presentations stand out.

As regards presenting your data within the covers of your thesis, that is something many of us struggle with; I know I did!