Post by robpintwala on Dec 2, 2015 20:23:45 GMT
Hey everyone,
In grad school I found that half of my time was spent fighting with (and losing to) various software tools. Granted, sometimes you want to obsess over the exact positioning of a legend on a graph in R. But most other times you just want to submit an abstract and go to bed.
The good news is that I discovered a number of really useful (and for the most part, powerful) tools that worked for me. The bad news is that I tried essentially everything available. I figure I'd spare someone else the trouble and share my preferences here.
Data Analysis and Statistics:
What I tried: SAS, SPSS, GraphPad Prism, Minitab, R, STATA, Statistica
What I liked: Minitab, R
Why?
Minitab is by-and-far the easiest statistics package to use, and is still fairly robust. It has virtually no learning curve, and the UI is simple and straightforward. Older versions produced terrible looking graphs, so I used something else for that. I would highly recommend anyone doing life sciences research work start here, if possible.
R is king. There's nothing it can't do, and if it's good enough for Facebook to use to analyse data on over a billion users, it's good enough for us. However, it's command-line only (you can install UI packages for it, but I found they were more cumbersome than helpful), so you should stay away if you're not prepared for an incredibly steep learning curve.
Graphs:
What I tried: Excel, GraphPad Prism, Minitab, R, SAS
What I liked: GraphPad Prism (liked doesn't begin to describe it, loved is a more apt term)
Why?
Graphs made in Excel tend to look like graphs made in Excel. Most statistics packages produce graphs that are relatively unmodifiable from stock (and getting your average stats package to produce a graph that will be accepted by the journal you are submitting to can be a struggle).
GraphPad Prism is stupidly simple and straightforward. It's clearly designed from the ground up to easily create beautiful graphs (enough so that I've had reviewers comment positively on them). Can't recommend it enough.
Reference Managers:
What I tried: EndNote, Mendeley
What I liked: Mendeley
Why?
This one comes down to personal preference more so than other tools, but I preferred Mendeley. It feels more modern than EndNote, and has handy features like in-browser citation saving (and .pdf importing). I'd say give Mendeley a try unless your entire lab / research group uses EndNote. Trying to coauthor a review paper with someone using EndNote resulted in manually managing some 100+ references, so just say no to collaboration. Or something.
Posters:
What I tried: Powerpoint, Adobe Illustrator, Poster Genius
What I liked: Illustrator
Why?
Vector art is obviously going to render better than raster art, especially when blown up for large format printing. But Illustrator also handles text and figure alignment better than any of the alternatives. Snap-into-place in Illustrator is never requires any fine-tuning or fighting (like in Powerpoint). Like all Adobe products there's a learning curve, but experience with other Adobe products like Photoshop or Lightroom will lead you to feel right at home.
If you want to keep it simple, though, Powerpoint works well enough for what it is. Poster Genius is an interesting idea (and good if you don't care to obsess about any detail of your poster), but is too simple in my opinion for producing a good poster.
---
Anyways, that's my take on some of the options you have when it comes to research work. Feel free to comment with your thoughts.
In grad school I found that half of my time was spent fighting with (and losing to) various software tools. Granted, sometimes you want to obsess over the exact positioning of a legend on a graph in R. But most other times you just want to submit an abstract and go to bed.
The good news is that I discovered a number of really useful (and for the most part, powerful) tools that worked for me. The bad news is that I tried essentially everything available. I figure I'd spare someone else the trouble and share my preferences here.
Data Analysis and Statistics:
What I tried: SAS, SPSS, GraphPad Prism, Minitab, R, STATA, Statistica
What I liked: Minitab, R
Why?
Minitab is by-and-far the easiest statistics package to use, and is still fairly robust. It has virtually no learning curve, and the UI is simple and straightforward. Older versions produced terrible looking graphs, so I used something else for that. I would highly recommend anyone doing life sciences research work start here, if possible.
R is king. There's nothing it can't do, and if it's good enough for Facebook to use to analyse data on over a billion users, it's good enough for us. However, it's command-line only (you can install UI packages for it, but I found they were more cumbersome than helpful), so you should stay away if you're not prepared for an incredibly steep learning curve.
Graphs:
What I tried: Excel, GraphPad Prism, Minitab, R, SAS
What I liked: GraphPad Prism (liked doesn't begin to describe it, loved is a more apt term)
Why?
Graphs made in Excel tend to look like graphs made in Excel. Most statistics packages produce graphs that are relatively unmodifiable from stock (and getting your average stats package to produce a graph that will be accepted by the journal you are submitting to can be a struggle).
GraphPad Prism is stupidly simple and straightforward. It's clearly designed from the ground up to easily create beautiful graphs (enough so that I've had reviewers comment positively on them). Can't recommend it enough.
Reference Managers:
What I tried: EndNote, Mendeley
What I liked: Mendeley
Why?
This one comes down to personal preference more so than other tools, but I preferred Mendeley. It feels more modern than EndNote, and has handy features like in-browser citation saving (and .pdf importing). I'd say give Mendeley a try unless your entire lab / research group uses EndNote. Trying to coauthor a review paper with someone using EndNote resulted in manually managing some 100+ references, so just say no to collaboration. Or something.
Posters:
What I tried: Powerpoint, Adobe Illustrator, Poster Genius
What I liked: Illustrator
Why?
Vector art is obviously going to render better than raster art, especially when blown up for large format printing. But Illustrator also handles text and figure alignment better than any of the alternatives. Snap-into-place in Illustrator is never requires any fine-tuning or fighting (like in Powerpoint). Like all Adobe products there's a learning curve, but experience with other Adobe products like Photoshop or Lightroom will lead you to feel right at home.
If you want to keep it simple, though, Powerpoint works well enough for what it is. Poster Genius is an interesting idea (and good if you don't care to obsess about any detail of your poster), but is too simple in my opinion for producing a good poster.
---
Anyways, that's my take on some of the options you have when it comes to research work. Feel free to comment with your thoughts.