Perhaps the proper way to present it is as an asymmetric error bar? That
is, treat "<0.05" as 0 (+ 0.05) (-0)?
Since Veusz supports independent positive and negative errors, it should
not be too difficult to create an input plugin which handles that
transformation.
Regards,
-- BKS
On 7/6/2014 11:30 AM, Henrik Rydberg wrote:
Thanks for your reply!
Well, I think your suggestions are good, but what I want to achieve is:
- presentation of the numeric value eventhough it is <0.05
- label or marker on those results so that the viewer knows it is below that
level of quantification
- ...but most important that an actual sample is taken and examined at that
level of quantification!
Hence:
NAN interpolate -> you wouldn´t know sample exist and is quantified
NAN skip -> as above
Treat as zero -> incorrect data, result could be up to limit of
quantification
Treat as max/min -> ?
Treat as marker -> maybe!
By using a split dataset in my example one could use set2 for plotting all
values and set1 as markers, that is "<" would be printed next to the
datapoint at the 0.05 level so that you could see:
1. sample is taken
2. value is verified to be below 0.05
3. we know that value is not 0.05, but denoted "<" and hence below limit of
quantification
regards
/Henrik
_______________________________________________
Veusz-discuss mailing list
Veusz-discuss@gna.org
https://mail.gna.org/listinfo/veusz-discuss
_______________________________________________
Veusz-discuss mailing list
Veusz-discuss@gna.org
https://mail.gna.org/listinfo/veusz-discuss
