Christian Zimmermann writes
> I emailed him already.
I email him prolly immediately when I saw it, and I think I was
up when it arrived.
> It has to do with when the server was down and I
> could not retrieve the data. I had a test for that, and it failed partially.
??? I don't remember it being down but
root@katri ~ # uptime
04:11:40 up 18 days, 9:59, 1 user, load average: 354.47, 354.33, 353.97
suggests it was. Anyway I am supposed to move to new server.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
L. Colin Xu writes
> I think there is likely a bug on computing "closeness measure" and
> "betweenness measure" in coauthorship network. I had always been in
> top 1% for many months, suddenly this month it is ranked around
> 20,000.
I think
http://collec.repec.org/nodes/x/u/2.html
has your current ranking at source in CollEc. It can change
every day.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
Reminder.
On Fri, 8 Feb 2019, Christian Zimmermann wrote:
> On Thu, 7 Feb 2019, Thomas Krichel wrote:
>
>> Christian Zimmermann writes
>>
>>> I am expecting to speak early May at a meeting of network economists.
>>
>> What meeting exactly? If I see what they work on, maybe I can see
>> if I have technology to offer.
>>
>
> A meeting of young economists working on networks. I expect them to have a
> sophisticated knowledge, thus saying you Dijkstra's algorithm is not
> sufficient. They will want details.
>
>>> I will talk about CollEc, and how this can be used to study
>>> networks. People likely will want to know some details about the
>>> computations. If you have any details I can share beyond what is on
>>> http://collec.repec.org/doc.html, they would be welcome.
>>
>> I could spruce up that document for the occasion. If you send me a
>> list of questions that would be helpful.
>>
>>> I will make a pitch for volunteers. Maybe I can find someone interested
>>> in
>>> helping with CollEc.
>>
>> Up for grabs if somebody wants to take it over. Rewriting it in
>> Python would not be hard because the web interfaces is
>> written in XSLT, so you don't have to redo that. In fact
>> it would probably be quite easy to redo the computations
>> in Python, and leave the search in Perl.
>>
>>
>> --
>>
>> Cheers,
>>
>> Thomas Krichel http://openlib.org/home/krichel
>> skype:thomaskrichel
>>
>
No connection to collec.repec.org at least for the last 12 hours.
Christian Zimmermann FIGUGEGL!
Economic Research
Federal Reserve Bank of St. Louis
P.O. Box 442
St. Louis MO 63166-0442 USA
https://ideas.repec.org/zimm/ @CZimm_economist
Christian Zimmermann writes
> I am expecting to speak early May at a meeting of network economists.
What meeting exactly? If I see what they work on, maybe I can see
if I have technology to offer.
> I will talk about CollEc, and how this can be used to study
> networks. People likely will want to know some details about the
> computations. If you have any details I can share beyond what is on
> http://collec.repec.org/doc.html, they would be welcome.
I could spruce up that document for the occasion. If you send me a
list of questions that would be helpful.
> I will make a pitch for volunteers. Maybe I can find someone interested in
> helping with CollEc.
Up for grabs if somebody wants to take it over. Rewriting it in
Python would not be hard because the web interfaces is
written in XSLT, so you don't have to redo that. In fact
it would probably be quite easy to redo the computations
in Python, and leave the search in Perl.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
Marina Azzimonti wrote:
> Alessandra Fogli, Veronica Guerrieri and I are organizing a "Women in
> Macro" conference. We did one last year and besides talking about
> economics, we like to have a short session to discuss gender issues in
> general. One topic that often arises is that women have a harder time
> "networker" than men do.
>
> One way to test this theory is to compare network characteristics of males
> and females. I noticed you guys at RePec (and Thomas at CollEC) have
> constructed a nice dataset which would allow to check this out. From what I
> see online, for each person on RePEC/Ideas you compute closeness and
> betweenness. Of course one would like to control for characteristics such
> as time since graduation, and perhaps sort by "quality". I see that it is
> possible to download a lot of this online, but I was wondering if you had
> an organized dataset already constructed you could share with me. This
> would save us a ton of time.
>
> What I think we would need is:
>
> - Name and Last name
> - Repec id
> - Affiliation
> - PhD year
> - Location (thta is, are they US or not, etc)
> - Network measures: closeness and betweenness (rank and value)
> - Gender
> - Various ranking measures
> - # of publications
> - fields
> - this is a stretch: list of coauthors
>
> I am not sure if there is any other information available, but this is what
> I see appears online.
>
> I could write a script to get the data from the site, but only the top 10%
> is available I believe.
>
> Anyways, let me know if this (or a subset of this) would be feasible. I
> think if we find something interesting it would be good to promote the
> ideas and CollEc datasets. Not many people knows about the network data and
> I can foresee a lot of interesting applications with it.
I can deliver the full CollEc data via rsync. I just set it up
krichel@trabbi~/collec$ rsync -va rsync://collec.repec.org/table collec.txt
receiving incremental file list
created directory collec.txt
./
CollEc.txt
sent 46 bytes received 8,097,988 bytes 2,313,724.00 bytes/sec
total size is 8,095,907 speedup is 1.00
This tabular data. It updated every day. I already deliver full
path data to Nikos. It's bulky
icanis@katri:~/icanis/paths/ras$ du -s biwe
140757804 biwe
If you make it to the city before Mach, we could meet. Or
if you can pay my LIRR tickets, I can come to Stony Brook
and give a talk about this work.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
Christian Zimmermann writes
> Site cannot be reached.
I know. I lost contact during an o/s update, but the sponsor
has put a two-factor authenticication to the Hetzer interface
making a reboot impossible.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
Christian Zimmermann writes
> The search function does not seem to work. No short-id is found.
I'm on the case. The search seems to run fine when the CGI
is invoqued on the command line.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
Matthias Weitzel writes
> I just came across the CollEC tool and played around with it for a bit.
> Unfortunately, when querying the shortest path to another author, the
> result is shown like this:
> [image: Inline image 1]
> I.e., there are only [] shown, I would assume the intended behavior would
> be to show names in those fields.
Christian Zimmermann writes
> The "search path to" button does not seem to yield any result.
I have seen this bug appear before, but it has never been clear to
me why it appears. Usually a restart of the icanis server---which
handles queries submitted to the web---fixes the issue.
I did an o/s upgrade on the server, and rebooted, it seeems
to have fixed.
BTW, the complete data, with all the paths is available if
you want it.
--
Cheers,
Thomas Krichel http://openlib.org/home/krichel
skype:thomaskrichel
