How Statwing (YC S12) Makes It Easier To Ask Questions About Data

I don't think it's for people who need SPSS. Statwing is great for much simpler use cases. When you need statistical heavy lifting, go to SPSS or R or something else that can do all that, but if you're not a statistician / data person and have simpler needs, things like Statwing (and, presumably, Wolfram Alpha Pro) can do a decent job.

Cofounder here. Totally agree that the hardest thing about data analysis is getting the data together.

But it's still true that once you have data, working with SPSS, R, etc. is a nightmare unless you're trained or somewhat technical. I actually did work with SPSS professionally for a couple years. And I was a lot like the roughly half of SPSS users who are nontechnical, and only need the basic analyses.

The two biggest painpoints for these users are knowing what test to run when and the difficulty of interpreting and working with the output. There's a ton of people, particularly more technical folks, for whom those are not big pain points, and that's fine too. They're just not in our target audience.

Also, RE the output comparison. Valid point if it looked like the implication was that SPSS can't output a histogram. Our point, though, is that we do it by default in the three clicks that it takes to output text, and it takes a lot more work to output even T-Test results in SPSS. Practically speaking, day-to-day nontechnical users don't visualize much in tools like SPSS because it's painful to do so.

Thanks for the comment, really appreciate it.

Or have the ability to connect to the local corporate Oracle database and get the data from there directly?

You mean like http://filepicker.io? :)

Appreciate that feedback. Sounds like getting data in is the theme of this whole page of discussion, too.

We agree. Hooking into existing sources of data will probably always need to be a big priority for us. Certainly not an easy challenge, but definitely solvable, bit by bit.

This is something that a favorite instructor of mine has stressed to me—the most difficult part of data analysis (and the part in which mistakes are most often made) is not choosing the correct analysis to do, nor actually performing it, but rather manipulating the data into the necessary shape/format. My favorite utilities for this are Hadley Wickham's plyr and reshape (http://had.co.nz/) (not sure if you already use these, figured I'd drop you a hint if not)

You might already know this, and it's not as automated as just dropping in a file, but one option for extracting tables from PDFs (other than retyping them) is the pro version of Adobe Acrobat. If you drag-select around a table, you can right-click to export it as Excel or CSV, which works most of the time.