Data Crunching and Charting....

[Mr BigSmoke]

Hi Tom.

I had more or less the same problem about 1 year and a half ago...
Since then i've tried either Matplotlib and chaco. I've started using
matplotlib but for some features reason i've decided to use only chaco.
Chaco is really cool but has some neg. points:
- documentation is not really detailed and there are really few
examples
- not supported anymore (at least i've never found any place supporting
it.. if anyone know bout any support tell me!!)
- sometimes it's really hard to look into it's code to extend some
classes...

The data i have to plot goes from very few curves and ticks to about
10/12 curves, 15000 ticks and range from -1 to 100000+ in Y axis.

Th plotting speed is good... but more data u plot, more problem u get!!
Like, zooming feature is not very "solid", and get's worse with tons of
data..., some features are missing (or at least i've never found them!!
I'd be very happy if someone says that i'm wrong and shoe me them!!),
like showing the values of th curvers when i pass the mouse on it, of
the highlighting a point in the curve if i clck on it!!

Fabio Pliger


tbonejo at gmail.com wrote:
> Hi all.
>
> Question: I have a project nearly complete written in VB.Net using
> charts from 3rd party vendors...expensive, yes, fast, not really. The
> data I am plotting is about 30 columns by 3000-9000 rows, all in a tab
> delimited file(43 files total). My question is whether or not I can
> accomplish the same in Python utilizing Matplotlib or Chaco for
> plotting, at a much faster speed. In my app it takes a while to loop
> through all the rows/columns to plot the data. Im just curious whether
> there may be a big boost in performance using python. Also I use
> spreadsheet style grids for displaying data. I was looking at the grid
> control for wxpython and that seems a nice fit. Anyone with experience
> with it have any opinions, Id appreciate them. Alot of the programming
> I do deals with crunching data and charting it. Sometimes I wonder
> whether python would be a much better canidate for this type of work
> rather than what Im using now.
> 
> 
> Thanks.
> 
> 
> Tom