Home Science CommunicationScholarly communication Graph theory: graphs are annoying

Graph theory: graphs are annoying

by Eva Amsen

I was away, but now I’m back. Report on trip later, after I’ve been a bit more productive. And there’s the problem.

I have spent all day afternoon evening fidgeting with stupid graphs. The graphs are in Excel, because that is the only program I have on my computer that makes graphs, and because the raw data were in Excel anyway, so I’m not changing that. Deal with it.

Just slapping the Excel graphs in my thesis won’t do, because they’re ugly, so I’m passing them through Illustrator, where I rewrite the labels (they say the same things, of course, but they look prettier) and remove all the nothingness. If you’ve ever pasted a graph from Excel into Illustrator you’ll know what I mean by “nothingness”, but I’ll try to explain for everyone else: about 50% of the image consists of just white boxes in the background of every single word or other piece of the graph. I don’t know why, it just does. You can’t even see that they’re gone because they were white (or even transparent) and behind everything else, but I’m taking them out because I’m an Illustrator purist.

Then the pretty gradient colours I used when I showed these graphs in a presentation are really ugly once they’re in Illustrator, and also not worth spending money on for the colour prints, so I’m replacing all the colours of the bars with just black.

And then WHOA, I noticed that the graphs are not the same size, so I had to go back to Excel to fix that part, and start all over.

The cat was sitting on my desk while I was doing this, and literally got sick. Not on the desk, which would have been easy to clean (minute mark 8:20 if you’re in a hurry, or 7:30 if you’ve got another minute, or why not watch the whole thing?)(by the way, I went to a Craig Ferguson taping while on vacation in LA. Harry Shearer was guest. Anyway…), but under the far end of the bed, near the wall, which is where the cat goes to puke. So there’s that to deal with.I have about 15 graphs, or maybe a million, I don’t know. I just did two, but they were just some _other_ graphs that aren’t even meant for my thesis. I don’t even know why I spent all that work on them. Clearly I’m going crazy!

I have about 15 graphs, or maybe a million, I don’t know. I just did two, but they were just some other graphs that aren’t even meant for my thesis. I don’t even know why I spent all that work on them. Clearly I’m going crazy!

Now I’m considering just saving the numbers and making new graphs from said numbers in lovely GraphPad at the lab tomorrow. Which means I can do something else now. Although I guess I already spent that time writing this.

And this is the first somewhat useful thing I’ve done since I’m back from vacation. I came back on Wednesday. Well, Tuesday night at midnight, so technically it really was Wednesday.

I was planning on handing this chapter in last Friday. It’s really almost done: it only needs the last few figures (consisting of a million graphs), the abstract, one paragraph in the intro, and some references looked up and cited. Why is this taking so long? Why is it so hard to get back on track?

update: I managed to prettify all my million (fourteen) graphs by just copying the data to GraphPad. I thought maybe I couldn’t use it because I already had the standard deviations, but there was an option to manually enter them. Yay!

Related Articles

13 comments

Bora Zivkovic August 18, 2008 - 2:13 AM

I have never used Excel for making graphs (I avoid using it for anything, unless forced by others). I keep reloading from computer to computer over the years the old CricketGraph III – it is simple, clean and does the job for the kinds of graphs I make.

Reply
Eva Amsen August 18, 2008 - 2:16 AM

I usually do use GraphPad, but the data were already in Excel so I thought it might be easy. (The program that calculates the RT-PCR times and converts them to levels of expression is in Excel. It’s a frustrating program, but I work with what we have, and nobody on the floor knows how to program anything.)

Reply
Eva Amsen August 18, 2008 - 3:05 AM

I downloaded the trial version of GraphPad and it made everything better! Woo!

Reply
Bora Zivkovic August 18, 2008 - 4:55 AM

I use GraphPad online calculator for almost all my stats.

Reply
Bob O'Hara August 18, 2008 - 5:18 AM

GUIs? Pah. Real statisticians draw their graphs from a command line. 🙂
I know a couple of guys who go 3 steps further: they edit their Postscript files (i.e. the file that gets sent to the printer) by hand.
Good luck with the net million. The Beast suggests that throwing up on/in shoes is optimal.

Reply
John Wilkins August 18, 2008 - 5:44 AM

Yes, GraphPad is your best solution, but really anything that isn’t by Microsoft is OK. MS Excel’s graphing features date from the time when it was all about building a raster, and so objects were easy to “erase” by putting a white box over them. Since the introduction of PostScript, you know, in *1984*!!!, that is unnecessary, but MS’s graphics routines haven’t been revised since. There’s a reason: the MS philosophy is “Never let them leave” – the Hotel California philosophy. So getting anything out of an Office environment, except to another MS product, is always troublesome.
If you have a Mac, try using Numbers. All graphics are in PDF already, so you can copy or open “print to PDF” files in Illustrator just fine. Cheap, simple and useful. Not much in the way of stats functions, but you already did that, right?

Reply
Henry Gee August 18, 2008 - 9:51 AM

_The cat was sitting on my desk while I was doing this, and literally got sick. … under the far end of the bed, near the wall, which is where the cat goes to puke. So there’s that to deal with._
My sympathies. I am reassured that this activity goes on in other places beside the Maison Des Girrafes.
!http://farm4.static.flickr.com/3254/2621226106_506616d1cd.jpg?v=0!
_Olympic Freestyle Projectile Vomiting Champion. Yesterday_.

Reply
Massimo Pinto August 18, 2008 - 1:16 PM

Is there anything free of charge that is good enough? I tend to avoid Excel too.
*Eva*: I think I am very interested in your RT-PCR code written for Excel. Should I pay a license on your PayPal account?

Reply
Eva Amsen August 18, 2008 - 2:12 PM

It’s not mine, it’s a BioRad macro in Excel. It takes values from the plate in one excel sheet and then reports the expression in the other.

Reply
Mark Tummers August 18, 2008 - 2:17 PM

I also use Illustrator for graph ‘beautification’.
Just make everything with white filling in excel, then copy paste. Then put in colours/gradients, and the like. Illustrator has all the tools you need to turn it into a sexy graph. Don’t stop with the simple changes. You can easily change the whole look of your graph until it looks like something worthy of Nature. Or even Cell.
Well, in theory. I never had a single graph in any of my papers, although I recently started doing them for more current projects.

Reply
Bob O'Hara August 18, 2008 - 5:00 PM

bq. Is there anything free of charge that is good enough? I tend to avoid Excel too.
There is OpenOffice, although they try to look too much like Micro$oft at times. If you’re prepared to learn the language, “R”:http://www.r-project.org/ is excellent: I use it for all my graphs now. But it’s not worth starting to learn it now, if you have to hand the chapter in last Friday.

Reply
Anna Kushnir August 18, 2008 - 6:53 PM

Excel makes me twitch. I did everything for my thesis in GraphPad. It didn’t help me understand statistics, but it certainly made for prettier graphs. I think it’s a lot more user-friendly than Illustrator, but with perhaps fewer options for graph beautification. Did you wind up prettifying the GraphPad figures in Illustrator anyway, or were they good enough?
Writing a thesis is stressful enough. You don’t need to make it worse by messing with Excel. Ick. It’s ok for software developers and professional statisticians, but not for bio researchers. Not ok by far.

Reply
Eva Amsen August 18, 2008 - 7:41 PM

The graphs are good enough now, I think. I made them so I can print those pages in cheap black-and-white, so they’re pretty plain.

Reply

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.