Re: Large analysis
- Hi all,
Did you know about the plugin for Excel which can analyse your own
I have just installed it. You need Microsoft Desktop Search for it
work, but after my laptop had indexed Outlook it produced a nice map.
Still not sure what to make of it though, but at least sign that
mapping is becoming much more mainstream. I found out about it from
the "Complexity and Social Networks Blog of the Institute for
Quantitative Social Science and the Program on Networked Governance,
Harvard University" (I wouldn't want to come up with an acronym for
On a quite separate note - but quite fascinating - UK newspaper The
Telegraph has available online SNA maps on interactions between
soccer players (ball passing) - see
You need to select a player and a match and then scroll to the bottom
of the page. Click on 'Pass./Rec' and you get an ONA/SNA map of the
Partner, Optimice Pty Ltd
www.optimice.com.au / www.onasurveys.com
--- In firstname.lastname@example.org, Peter Biddle <peternbiddle@...>
>Valdis' questions. They aren't so automated yet that you can just
> There are email tools which can help answer at least some of
press one magic button and get all the answers, but that is the
direction folks seem to be headed.
> What I think you want is an email analysis tool which can tell you:
> Who emails whom?
> How often do they communicate?
> Is the communication one-sided or evenly distributed?
> What are people mostly talking about?
> Who communicates in those areas (so focus on topics in addition to
> Where do new ideas tend to originate?with email (and eventually IM, phone logs, chats, wikis, blogs) for
> Where do new ideas tend to migrate to?
> Combining email with a survey seems like a great solution. Start
everyone in a given network. With an appropriate opt-in privacy
system, of course.
>the org - pivotal people (e.g. brokers) as well as those who have
> Then use a rough analysis of that data to identify key facets of
somehow wound up isolated or displaced (eg orphaned individuals or
groups). We could then use the resulting data to target the
recipients of a survey as well as to tune the questions themselves.
Once the survey is answered, we could apply our learnings from that
back to the email analysis to see where our assumptions held out and
where they need to be adjusted.
>and hopefully the company undertakes changes. Ongoing email analysis
> Presumably some recommendations will come out of all this analysis,
can help mgmt see if the changes they choose to take are shifting
communication flows and improving things, allowing them to tune
>analysis of email. I happen to run the dev team for one of them so I
> There are companies out there which provide tools to do automated
won't mention any company here as it may be perceived as marketing.
But if you are interested, look around!
>0400Subject: Re: [ona-prac] Re: Large analysis
> To: ona-prac@...: valdis@...: Wed, 25 Jun 2008 19:25:23 -
>management wants to know specific things and therefore ask specific
> Yeah, proxies are fine if they help answer your question. But often
questions. Proxies can be use their as a "reality test" but are
insufficient on their own.A specific proxy can be defined as in this
project.http://www.orgnet.com/email.htmlWe looked at emails just for
particular project members, allowing us to get a pretty good proxy
for "overall project communication". Yet, we did not know who was
talking tech, who was talking admin, who was talking process, who was
coming up with ideas, or who the experts were.There is no free lunch
nor simple swaps in SNA/ONA.ValdisOn Jun 25, 2008, at 6:40 PM,
jkbuenen wrote:> Valdis, Giancarlo,>> I know I probably butt in far
to early being a freshman on this group>> How about using the
companies computer mediated communication (cmc) > data as a> proxy?>
Our limited experience, and some scientific work, shows that there >
is a good chance cmc> can be taken as a proxy. Off course you always
have the possibility > of doing a limited> survey based analysis to
validate the proxy for this particular case.> We also accept working
as sub-contractors in this field.>> Our software allows a dynamic
picture with role developments showing > over time.>> Appreciate your
feedback>> Kind regards,> Jan-Kees> www.centerconsult.net>>>>>> ---
In email@example.com, Valdis Krebs <valdis@> wrote:>>>>
Giancarlo,>>>> You CAN do a roster survey of 400... divide them up by
department/>> location and then alphabetically within the dept. We
have done this>> and it works OK... allowing people to quickly find
those natural>> groups/locations they interact with. IBM has done
populations of>> 600-700 using this method.>>>> Your response rate
will determine which metrics will make sense for>> you... don't
choose flow betweenness, or anything else, before you>> have a ready
data set! Also, are you asking symmetric or asymmetric>> questions?
>>>> Valdis>>>>>> On Jun 25, 2008, at 12:07 PM, Giancarlo Orianiwrote:>>>>> Dear friends,>>> a large companies is about to analyse
its information flows in order>>> to identify the informal
information brokers. The company include>>> some 400 persons. A
roster technique approach is clearly impossible.>>> I would suggest
them a free recall appoach, asking each respondents>>> to select 30
actors and to evaluate then their links to the 30. Then>>> I will
calculate the flow betweenness, even if I know the results>>> won't
be perfect.>>> Do you know a better method?>>> Thank you so much>>>
Giancarlo>>>>>>>>>> ------------------------------------>> Yahoo!