In Tallin three groups met and discussed three themes, each of which is briefly reported on below.

Twitter/ Flickr Paper and continuation

Objective: How to proceed with analysis of Twitter/Flickr data on London
Objective of the study: Find distinctive places in geo-social media
Focus first on (qualitative) analysis of the actual terms used, from a small, purposeful sample (in other words: Look at actual lists of terms, not binary term vectors and derived statistics)
Compare similarity between Twitter/Flickr places aggregated using administrative / demographic boundaries
Compare with geo-demographic characteristics

VGI Ontology

This activity aims at building an ontology of VGI concepts. Implementation by Gilles and Rob with help of a support group of ENERGIC members.

Common understanding of VGI concepts within ENERGIC
Create applications and database schema
Usage of ontology of tasks for evaluating of data quality
Relationships with other domains
Semantic enrichment (related to bookchapter ‘VGI Enquiring’)

Keyword extraction of ENERGIC book text
Conceptualisation and grouping of concepts
Mapping with exisiting ontologies (Dbpedia, Wordnet, Geoname, etc)
Visualisation of ontology
Request for input from ENERGIC members – feedback to 2.

Formal ontology & visual representation in ENERGIC repository (wiki style versioning)

STSM Sept-Oct 2015 -> MCM activity Oct-Nov 2015 -> Publication spring 2016


Organize a hackathon in multiple locations around Europe
Input: list of VGI sources and methods to access their data
Output: extract, analyze and/or visualize VGI from the above defined sources
Possible difficulties:
#1: focus on data extraction and/or data analysis
#2: predefine the list of data sources or specify in advance
#3: combined teams (computer science + geographer) or non-expert volunteers?
Possible themes:
Evaluate landscapes
Evaluate a success of a city or multiple cities based on the collected VGI
Quality of life
Public transport – not clear how and what?
Possible analytical dimensions:
Temporal analysis of different VGI sources
User profiling based on different VGI sources (nationality, location, sex, etc.)