Max Spevack wrote:
> On Wed, 18 Mar 2009, Yaakov Nemoy wrote:
>
>
>> i'm gonna blog this, but here's the latest output i get from scanner
>>
>>
http://ynemoy.fedorapeople.org/ekg/graph.test.003.html
>>
>
> >From a raw data perspective (which is what I asked you to focus on),
> this is fantastic!
>
> --Max
> _______________________________________________
> ekg mailing list
> ekg(a)lists.fedorahosted.org
>
https://fedorahosted.org/mailman/listinfo/ekg
>
Sweet, good to see. It would be nice to see totals mixed in with the raw
data per domain for each month. Also, in addition to totals, inside vs
outside totals... what's the %redhat-owned-domains look like?
Another tip, and I know this is just rough, but if you are using hyphen
as the delimiter, do negative numbers like (12) rather than "-12", and
possibly also show them in Red. I imagine having
report modes in both formats is equally important. (Ideally an interface
that allows us to see an index of all lists and then click on one in
either graph or number format, or pick two or more to compare in numeric
format would be ideal)
We need the numbers as much as the graphs, sure enough. What you have
there is not a huge leap to make it do a line of changing pacmen :)
Another idea -- You mentioned the kevin bacon stuff in your blog post --
I really don't think that's all that interesting (having done it), but
it might be cool to apply the github graph concept to see what a
person's posting mix is over time, does he post more on certain lists
and how that changes. This is less useful data, but I think it might be
neat to see. It could possibly be applied to a list of domains as a set
instead of just a person, so we could watch for a given set of domains
and a set of lists the participation in them rise and fall.
For instance:
List set: all systems management apps
Domain set:
redhat.com,
gmail.com, others
Format: github impact graph style
So we see that over time redhat shifts from working on FOO to BAR, but
gmail was originally more interested in BAZ and is gaining interest in
FOO, but other is interested in BAZ originally and then moves interest
to FOO.
The idea here is ... is our involvement in communities of a certain type
the same as our own communities involvement, or do we have different
alignments?
I'm not sure sure of the best way to look at this numerically. At some point i'm
going to look at how to aggregate certain domains under organisations and do an analysis
from there. You'll be able to see this as a raw number report pretty much right
away.
As for the technology side of things, i love the github widget you found. I spent some
time this weekend looking through it, and as i said before, i want to package that up for
further use. I think applying the data to it shouldn't be too difficult, it's just
a matter of how much time am i going to focus on reports in the next two weeks and how
much time am i going to focus on making it look pretty.
Also on a technology side of things, i'm going to explore looking at aggregating
different mailing lists / other streams of data under certain projects and/or topics, and
seeing what kind of analysis we can do there. I'm taking the approach that every
function and method i write should be a general purpose analysis function for a
generalized dataset. It's something that's on the agenda for sure.
-Yaakov