- CollEc-run - example.com

Re: [CollEc] mentioning co-authorship registration failure in CollEc
by Thomas Krichel 09 Nov '19

09 Nov '19

Danny Cassimon writes > Apologies for contacting you to mention a concrete broken link or > failure of account of the CollEc part of the REPEC/IDEAS repository. We don't get enough of such feedback! > As an experienced user of the system, I now note that recently, the > system seems to fail to, or is in any case much slower in, accounting > to register new co-authors in personal co-author lists. To give a > personal example, a colleague that registered about two months ago, > and with whom I have a mutually-accepted co-authored paper (see: > https://ideas.repec.org/p/pra/mprapa/96072.html) Are you talking about Solomon here? Here is the input directory icanis@helos:~/icanis/input$ ls -lt total 388168 -rw-rw-r-- 1 icanis icanis 70223531 Nov 8 04:10 ras_mans_texts_1573182608.xml -rw-rw-r-- 1 icanis icanis 8624782 Nov 8 04:10 ras_mans_edges_1573182608.xml -rw-rw-r-- 1 icanis icanis 70223531 Nov 7 04:07 ras_mans_texts_1573096074.xml -rw-rw-r-- 1 icanis icanis 8624784 Nov 7 04:07 ras_mans_edges_1573096074.xml -rw-rw-r-- 1 icanis icanis 70223531 Nov 6 04:07 ras_mans_texts_1573009656.xml -rw-rw-r-- 1 icanis icanis 3218932 Nov 6 04:07 ras_mans_nodes_1573009656.xml -rw-rw-r-- 1 icanis icanis 8624786 Nov 6 04:07 ras_mans_edges_1573009656.xml -rw-rw-r-- 1 icanis icanis 70223531 Nov 5 04:05 ras_mans_texts_1572923121.xml -rw-rw-r-- 1 icanis icanis 8624771 Nov 5 04:05 ras_mans_edges_1572923121.xml -rw-rw-r-- 1 icanis icanis 70223531 Nov 4 04:08 ras_mans_texts_1572836929.xml -rw-rw-r-- 1 icanis icanis 8624778 Nov 4 04:08 ras_mans_edges_1572836929.xml The edges matrices don't seem to change too much. icanis@helos:~/icanis/input$ grep psa1752 ras_mans_edges_* icanis@helos:~/icanis/input$ Solomon is not there. Danny is icanis@helos:~/icanis/input$ grep pca462 ras_mans_edges_1573182608.xml <edge from="pbl118" to="pca462" length="0.666666666666667"/> <edge from="pca1004" to="pca462" length="1"/> <edge from="pca462" to="pcl16" length="0.4"/> <edge from="pca462" to="pde1008" length="2"/> <edge from="pca462" to="pdr61" length="0.2"/> <edge from="pca462" to="pen79" length="0.12"/> <edge from="pca462" to="per126" length="0.5"/> <edge from="pca462" to="pes101" length="0.0769230769230769"/> <edge from="pca462" to="pfa375" length="1"/> <edge from="pca462" to="pfe372" length="1.5"/> <edge from="pca462" to="pme555" length="0.166666666666667"/> <edge from="pca462" to="pme591" length="3"/> <edge from="pca462" to="pra263" length="0.272727272727273"/> <edge from="pca462" to="pri349" length="1"/> <edge from="pca462" to="pse638" length="0.666666666666667"/> <edge from="pca462" to="ptc17" length="0.333333333333333"/> <edge from="pca462" to="pva229" length="0.109090909090909"/> <edge from="pca462" to="pva653" length="0.545454545454546"/> <edge from="pca462" to="pva685" length="2"/> <edge from="pca462" to="pve127" length="2"/> <edge from="pca462" to="pve344" length="0.25"/> <edge from="pca462" to="pwo109" length="1"/> icanis@helos:~/icanis/input$ The input update is obviously broken. The input script uses icanis@helos:~$ grep per_dir perl/compose_ras_icanis_input my $per_dir="/home/adrepec/RePEc/remo/per/pers/"; This is NFS linked to helos on aigtu icanis@helos:~$ ls -ld /home/adrepec lrwxrwxrwx 1 root root 25 May 25 2017 /home/adrepec -> ../mnt/aigtu/home/adrepec There are psas in there icanis@helos:~$ ls /home/adrepec/RePEc/remo/per/pers/s/psa17* /home/adrepec/RePEc/remo/per/pers/s/psa1700.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1701.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1702.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1703.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1704.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1705.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1706.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1707.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1708.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1709.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1710.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1711.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1712.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1713.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1714.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1715.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1716.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1717.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1718.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1720.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1721.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1722.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1723.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1724.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1725.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1726.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1727.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1728.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1729.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1730.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1731.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1732.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1733.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1734.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1735.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1736.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1737.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1738.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1739.rdf /home/adrepec/RePEc/remo/per/pers/s/psa173.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1740.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1741.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1742.rdf /home/adrepec/RePEc/remo/per/pers/s/psa1743.rdf /home/adrepec/RePEc/remo/per/pers/s/psa177.rdf But no psa1752 icanis@helos:~$ ls /home/adrepec/RePEc/remo/per/pers/s/psa1752* So CollEc is not to blame, the input is missing. -- Cheers, Thomas Krichel http://openlib.org/home/krichel skype:thomaskrichel

2 1

katri --> helos
by Thomas Krichel 07 Aug '19

07 Aug '19

This is finally finished. The new server is an Hetzner EX-52. There is no Debian for this hardware yet. I went with Ubuntu. Some stats root@helos ~ # free total used free shared buff/cache available Mem: 65877048 2862536 12860392 2348 50154120 62276476 Swap: 33521660 0 33521660 root@helos ~ # df | grep md2 /dev/md2 7717425284 1773026772 5555391392 25% / I run 9 update scripts on amorp (via NFS) and 9 on helos. Maybe I should run more on helos. I put together a stats script for the path files. count: 42994 age: 0.000497685185185185 51.1333999371576 79.2213541666667 This last line has min, mean, and max ago of files in days. I am shocked about the count. Is this correct? Can we really have close to 43k in our largest component? -- Cheers, Thomas Krichel http://openlib.org/home/krichel skype:thomaskrichel

2 4

Inconsistent data
by Christian Zimmermann 04 Jul '19

04 Jul '19

There is a problem either in the file at ftp://ftp.repec.org/CollEc/CollEc.txt or with what is dispayed on the site. The numbers do not match. For example for me: site: closeness 4943 4.8307 betweenness 918 7.87024 file: 4.98787 5804 7.25581 937 The difference is larger for people that are complaining. 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

2 1

Incomplete data
by Christian Zimmermann 09 May '19

09 May '19

The ftp://ftp.repec.org/CollEc/CollEc.txt file is incomplete! It stops at pma1077. 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

2 9

Re: [CollEc] potential errors
by Thomas Krichel 07 Apr '19

07 Apr '19

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

1 0

Re: [CollEc] potential errors
by Thomas Krichel 07 Apr '19

07 Apr '19

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

2 3

Re: [CollEc] [External] Re: CollEc
by Christian Zimmermann 05 Apr '19

05 Apr '19

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 >> >

2 1

site down
by Christian Zimmermann 19 Mar '19

19 Mar '19

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

2 1

Re: [CollEc] CollEc
by Thomas Krichel 08 Feb '19

08 Feb '19

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

2 1

Re: [CollEc] Network Dataset
by Thomas Krichel 18 Jan '19

18 Jan '19

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

1 0