Agenda 5-1-15
- Schedule week:
* invert order speakers Mon 12-1 (Karin last)
* roadbook with roles / can be also tabliers at the wall:
performing continuity: organising the day, make sure no one is lost, tell & re-tell, updating people
expert: for 2nd week, a tour role to make sure someone invites non-programmers to help, can be one dedicated person per half day, for Python, git...
metaphorist: for 2nd week, a tour role to make sure someone invites non-techs for a non-tech session/work group on some algorithmic procedures...
documenter: someone taking on the task to write a blogpost about the highlights of the day, and keep an eye on documentation platform, that everyone leaves traces of what they do on CONSTANT SERVER, add to the '100 ideas list'
kitchen crew: for lunch we need people who bring up/down 3 extra tables from the basement, who can help set the tables, clean etc - not on 15 and 23 Jan
- Look at documentation platform: notes, links to datasets/references, images, videos, scripts
* organise a 'sharing policy'? commenting your code, describe your script
- mail to mailinglist: program of the weeks, invitation to introduce themselves + bring datasets/texts (specify)
-> introduce ourselves to mailinglist
- Check budget: foresee budget + time for post-documentation / publication
Agenda 11-12-14
with Karin, Anne-Laure, Martino, Nicolas, An
10-11h: WEEK 1: Data-portraits of a potato* introduction Karin
* discussion on content/timing lectures Mon 12-1
* preparation of material: datasets, texts/reader, video clips, links to potato info, tools (who? team? planning? see spreadsheet planning)
* preparation of workshops: Open Office, R, manual practises (who? team? planning? see spreadsheet planning)
11-12h:
* latest news on guests (ao workshop CliPS)
* invitations: add latest confirmations, mail addresses, check budget transport/lodging
* - Nova (Anne-laure), Denis (A), Open Data (A), Pinkscreens (AL), slug (M), hackerspace.be (Bssl/Gent) (A), BAM, Mediawiki, /etc, Donestech, Debianwomen, ....
* - nY (lit magazine, A), nettime fr/nl (Martino), netbehaviour (A), spectre (N), ...
* proposal An for task division (preparation/worksession Week 1/2)
* next meeting 5th Jan 15h-18h
TODO:
* install wordpress (Nicolas, An): 17-12?
* create mailinglist Cqrrelations (Nicolas?)
* provide good working Etherpad for documentation (An checks with Femke)
* create user 'cqrrelator' on Owncloud & share folder (An)
* feed hotglue (Martino, Peter): 5-1?
* install network in deBuren, shared hard drive/Piratebox (Michael 8-1)
*
Agenda 3-12-14
with Anne-Laure
0. introduce Anne-Laure to team/program
1. Overview sessions:
* week 1: data-portraits of a potato
- live: Karin Ulmer (context/strategies/responses), Aprodev
I have been looking into some data, and can prepare a sheet for potatoes - beyond trade, also looking into production figures, some legislative bits, protected names of potatoes, hectares used, yields attained, seeds trade, etc.
I can give it a try to check more data on the EU export helpdesk. What kind of questions do you have? Depending on the questions, I could either try to figures them out myself, ask somebody, or encourage you to write to the Commission directly, they have a helpdesk staff support. You could present yourself as you like, or as a potential exporter...
Anne-Laure (statistics/tools?) present software tool R (en ligne de commande/libraries spéciales) + Excel/Open office?
bon statisticien: possible de mettre en lumière limite des données, mais comment réalité est avalé par la nomenclature 'tu veux que tout soit dans les données', 'on veut toujours plus de colonnes!', est utopie, il faudra toujours discuter avec les gens
effrayant: le coût de cette maintenance de nomenclature pas supérieur au reste? 500 pages de nomenclature sur le site de l'inami; bases de données spéciales pou rles implantations, pacemakers ave cle no d'appareil - s'il y a problème technique, il faut retrouver tous les patients
-> institute of noenclature ex ATC: pour les médicaments
statistique est outil de pouvoir, possible de produire arguments autonomes
faire mal le datamining = pas un risque, c'est une question de compétence
question de la couche de nomenclature et 'données sur le monde' et les moyens qu'on met à ça, est délirant, inflation dans nomenclature/précision - information est utilisée pour cibler
'shit is in, shit is out': informations mal adaptées à ta question, ne va pas te donner la réponse, problème de qualité de l'analyse ex. données ne disent rien sur 'la qualité de travail' de la personne
mauvais statistique vs danger de la statistique
different problems of power: use data to relate to life & tools for dealing with these data
le fait que les données sont construites, it is always a bias
there is no neutral data, if you're a good analyst you know...
ex. France: pas de données sur la propriété (évalué sur base de revenu du patrimoine, mais pas par rapport à la valeur de l'objet), patrimoine n'est pas nécessairement revenu
ex. oui, chiffres sur le production (BNP), taxes (TVA)
-> contact people of Open Data (also in Gent, Dimitry Mediawiki)
Joost Colpaert (B-rail), porte le projet Open Data
Valery Vermeulen (tools/context stockmarket), Stéfan Marsiske (Parltrack, tbc), Michel Buisson (agro-economist, data & agriculture)
- skype: Geoff Cox (approaches), Maria Pttq (approaches?), Anne Helmond
(theory?)
-> only technical expertise of Anne-Laure and evt Stefan Marsiske for those who would like to experiment with datavisualisation/mining -> Al sends invitation to her former professor
-> find 2nd person?
nomenclature (code & définition du code) -> how is it set-up? who decided on classification?
-> description, vérifié avec entreprises... régulièrement adapté, changement = cauchemar
travaillent entre langue & la loi ---> produit plein d'information
'nuts': permet comparer régions, provinces, communes en catégorisations
cfr VJ9, p. 128: list of names
implique administration pour aller négocier avec les utilisateurs, aussi formalisé et malléable comme la loi
AL: 'le grand problème c'est que la langue n'est pas une nomenclature', qu'est-ce que tu appelles un 'employé'? pour pouvoir parler de l'histoire, il faut savoir quantifier
-> données/nomenclature forcent résumé de la réalité, mais en même temps ils fournissent de l'information... comment maltraiter les données
OUTILS:
http://exporthelp.europa.eu
Source: Stat-activisme - Comment lutter avec des nombres (Isabelle Bruno, Emmanuel Didier...), Editions La Découverte, Paris, 2014
* 2e semaine:
- live: Guy De Pauw (workshop classifiers Python/clips), Michael & Nicolas (workshop on scripts developed in Scandinavia??)
- texts to work with: some about potato & law, crawling on tweets etc, NL/EN of 1984 (j'ai les droitsI have rights till 2016), other public domain texts
- cfr tools/methods CLIPS/Doc
2. what can we learn from Genderblending, prepare worksessions
structured but vague enough
enough participants & structure they can react to
check with Femke
3. look at invitations: deadline 15-12
-> send invitations to mailinglists (Open data, writers)/agenda, send reminder with deadline
4. hotglue/wordpress (next week?)
5. welcome pack
-> Alex & Steph: kindergarden/babysittting?
-> next meeting 11-12 at 10h (Anne-Laure, Nicolas, Martino, An)
check separate meeting for Patterns, 11-12 in namiddag
fix a day beginning of January / separate tasks?
Agenda 26-11-14
* invitations - send out tonight/answer before next meeting + public call
* overview sessions
practical info & suggestions
* overview events http://note.pad.constantvzw.org/public_pad/Cqrrelations_public
Recyclart -> prepare it in hotglue, anounce informal meeting in the invitation
Preparation content w/ participants
* public web page: wordpress - nicolas/martino
* translations FR for newsletter - nicolas/an
* hotglue - all add links
-> next meeting 3-12 at 17h (Peter, Martino, An) + 11-12 at 10h (Nicolas, Martino, An)
- what can we learn from Genderblending, prepare worksessions
- look at invitations
- hotglue
- welcome pack