[R-sig-ins] General Insurance Data

[David Winsemius]

I was puzzled when I read that this program was "private". It seems to be a violation of the licensing terms of R.  It appears to depend entirely on the functionality of R which would make it a derivative work.

-- 
David Winsemius
Sent from my iPhone

> On Oct 1, 2014, at 1:22 PM, Dan Murphy <chiefmurphy at gmail.com> wrote:
> 
> Thank you for that reference. Looks nice. More than the original poster was
> looking for?
> 
> IMO, open-source is a great vehicle for distributing the guts engineering
> that accommodates data exchange and munging. OTOH, I'm not convinced that
> open-source is the right vehicle for distributing the "look and feel" of a
> GUI. But I'm "open" for persuasion. :)
> 
> Dan
> 
> On Tue, Sep 30, 2014 at 10:37 AM, Brian Fannin <BFannin at redwoodsgroup.com>
> wrote:
> 
>> Can’t believe I forgot to mention that! I remember the presentation and
>> it looks to be a very powerful tool. The integration with LaTeX is
>> particularly nice.
>> 
>> 
>> 
>> *From:* Christophe Dutang [mailto:dutangc at gmail.com]
>> *Sent:* Tuesday, September 30, 2014 11:30 AM
>> *To:* Brian Fannin
>> *Cc:* Edward Roche; Markus Gesmann; chiefmurphy at gmail.com;
>> R-Sig-Insurance at R-Project. Org
>> 
>> *Subject:* Re: [R-sig-ins] General Insurance Data
>> 
>> 
>> 
>> On the GUI part, there is a private software RPGM (
>> http://www.pgm-solutions.com/) offering what Edward is looking for.
>> 
>> 
>> 
>>  This tool was presented at the r in insurance conference this year also!
>> 
>> 
>> 
>>  It is partially developed by an actuary.
>> 
>> Christophe
>> 
>> 
>> Le 30 sept. 2014 à 15:25, Brian Fannin <BFannin at redwoodsgroup.com> a
>> écrit :
>> 
>> Edward,
>> 
>> 
>> 
>> If I’m reading your message right, the problem seems to break down into
>> two parts: 1) reshape claims data, either by aggregating individual claim
>> amounts or reshaping data from wide to long. 2) Provide a GUI to guide the
>> user through the process of munging the data so that it may be analyzed.
>> 
>> 
>> 
>> On the first point, it would be helpful to identify a typical structure
>> which must be modified. MRMR doesn’t have a function to go from wide to
>> long as that’s a very simple melt command. The data I use in my work lives
>> in a RDBMS in long format, so I’m sorted. I’ve toyed with the idea of a
>> “collapse” function that would aggregate data along temporal dimensions, or
>> summarize across a hierarchical axis (i.e. sum all of the territories into
>> a single country). Such a function would be general enough that it could
>> summarize individual claim transactions, though that’s likely better done
>> by a database server.
>> 
>> 
>> 
>> One the second point, note that an absence of a G