Ask HN: Tool that learns table structure and produces summary?

6 points by dkural 12y ago ↗ HN

Hi Folks,

What tools do you know of out there that can "learn" what's in a CSV / TSV file, make some smart guesses about the columns, and provide some sensible metrics / summary / plots of the data?

I know Google Refine, but it doesn't quite fit the purpose. What other stuff is out there?

4 comments

[ 5.0 ms ] story [ 16.8 ms ] thread

vittore 12y ago ↗

SSIS can do it (suggest columns feature). Also from MS SQL MS you can do the same through the "Import Data" which basically creates one time SSIS package.

gtirloni 12y ago ↗

Perhaps you can learn something from this:

http://kimonify.kimonolabs.com/kimload?url=http%3A%2F%2Fwww....

dkural 12y ago ↗

This is pretty amazing. Yes, this is in the right direction :) In a way what I have in mind is simpler: The data is already more structured, but since it is a big blob of rows, the user has no idea of relative proportions of things in it and needs to spend a lot of time querying, plotting etc.

So the steps would be: #1 Recognize the data types sensibly (Kimono-like) #2 Size-up the data, give user some feel for the overall content.

Any tools that do #2?

tzm 12y ago ↗

I wrote Spreadsheet.io that is similar to Google Refine (and Fusion Tables) and performs various data type analysis for visualization, etc. There's also a native Excel add-in that enables a Javascript API bridge to an existing sheet. Working on a release.. it's been a side project.