Past Meetup

Open Data and Natural Language Processing

This Meetup is past

76 people went

Radialpoint

2050 De Bluery, Suite 220 (2nd floor) · Montreal, QC

How to find us

If you cannot get in after 6pm, call Matthieu: +1-438-820-7128

Location image of event venue

Details

Open data and Natural Language Processing

Happy New Year! We’d like to kick off our Meetup series for 2015 with two very interesting topics. Leading up to Open Data Day and the CODE 2015 Hackathon to be held from Feb 20-22, our first theme for this meetup is Open Data. Our second theme for this meetup is Natural Language Processing.

Our first speaker will be Stephane Guidoin (https://www.linkedin.com/in/stephaneguidoin) from OpenNorth (http://opennorth.ca/about/) in a talk titled - La standardisation, un mal nécessaire? (Standaridization, a necessary evil?) Our second talk will take a bold look at an application of Natural Language Processing (NLP). Our speaker will be Pablo Duboue (http://duboue.net/), a practicing researcher in NLP. Lastly, we'd like to get you warmed up for CODE 2015! Join us in a Yack+Hack session where we'll be navigating Real Estate data made available from http://open.canada.ca/ (http://open.canada.ca/.). We'll put up some initial code, and aspects of the data sets we find interesting. We would like you to take a look at the code, modify it if you like or meet us at our Yack+Hack to literally just yack, discuss and hack!

Agenda

6:00 PM - Doors open, get some beer and pizza. Provided by RadialPoint.

6:30 PM - Stephane Guidoin - La standardisation, un mal nécessaire? (Talk in French)

7:00 PM - Pablo Duboue - Natural Language Processing (Talk in English)

7:30 PM - Break

7:45 PM - Yack + Hack Session - Navigating Real Estate Open Data Sets with Vaughn DiMarco

Abstract for the talks

Stephane Guidoin - La standardisation, un mal nécessaire?

FR: L'ouverture des données est une chose. La mise à disposition de données de qualité et pouvant être utilisées à travers plusieurs juridiction en est une autres. Avec l'augmentation du nombre de gouvernements adoptant les données ouvertes, la nécessité d'une certaine forme de standardisation des données devient nécessaire. À l'aide d'étude de cas rapide, Nord Ouvert présentera sa vision des choses ainsi que son implication au sein de l'Open Government Partnership.

Pablo Duboue - Natural Language Generation for Statistical Reports

Most times, data visualization covers only graphical methods to present the data. Interestingly, human written statistical reports seldom rely exclusively in graphics; they involve a mixture of text and images. In this talk, I will summarize classical work on natural language generation for statistical reports [1] giving a short tutorial of OpenSchema [2] with SimpleNLG [3].

To conclude, I will discuss my ongoing Thoughtland project [4].

[1] For example, the Postgraphe system, developed at UdeM
http://www.iro.umontreal.ca/~fasciano/recherche/postgraphe/article-INLG96.pdf
[2] https://sourceforge.net/projects/openschema/
[3] https://code.google.com/p/simplenlg/
[4] https://github.com/DrDub/Thoughtland