-
Notifications
You must be signed in to change notification settings - Fork 1
Status Report
Hannes Mühleisen edited this page Jul 12, 2016
·
11 revisions
- First steps: Build simple Twitter Bot, Literature Research, Data sources for activities in entire NL
- Example code: https://github.com/hannesmuehleisen/fotofalle
- MonetDB node connector https://www.npmjs.com/package/monetdb
- Next Meeting on Skype, Thursday 9:00 hannes.muehleisen
- APIs
- Ahmad’s idea: extract from Facebook <= could be nice in general
- feedback to our bot - without asking for it
- ask explicitly
- check where following tweets did come from
tourist detection algorithm <= something on github?
- location where they come from?
- check location of last tweets
- analyse language
- Get basic bot to run
- Hannes is looking for office space?!
- Ahmad is setting up a twitter profile for the bot
- Maartje & Ahmad will add a short summary to the CroM-SoM documentation (where did the illuminati go?)
- Maartje and Ahmad looked at the Twitter API. Result: Easy.
- Some short literature research, no results yet.
- Kathrin found workspace at DataLab on Wesperstraat
- Are tourists actually posting on Social Media while in AMS?
- How to detect if someone is a tourist (Profile information, previous posts)
- Setting up Twitter profile for bot (Ahmad)
- Doing more literature research (Pro Googling)
- Making slides for Thursday ADS meeting, 9:30 meeting in Kathrin's office on Thursday at VU
- Start saving tweets/instagrams from a 5km circle around Dam square, store on cromsom.nl box
- ADS Presentation Thursday
- ADS on Thursday (June 23) (Meeting before at 9:30 in Kathrin's Office at VU)
- Skype-Meeting (June 30, 16:30 Amsterdam Time)
- Maartje: Use tracking beacons with low-energy bluetooth to track where people are actually going? Issue is that we need own application and need bluetooth all the time.
- Ahmad: First local tweet Viz at http://cromsom.nl:8080/ - nice!
- ADS Presentation: Related project "Buurt in zicht" Put slides into repo
- Save tweets to persistent disk, make sure all of JSON is there (either flat files, or document store)
- Ahmad: Figuring out semantics behind Twitter JSON
- Maartje: Literature review, looking at hackathons, companies
- Follow-up http://www.nrc.nl/handelsblad/2016/04/15/slimme-stad-of-dataslurper-1609291 Email Rutger (Kathrin)
- 2016-07-07 16:00 at CWI
- Maartje: Literature research done for now, found various related papers (to go into git repo)
- Amsterdam Institute of Metropolitan Solutions is a possible cooperation partner
- Ahmad: Twitter data available at
cromsom.nl:/home/ahmad/logs
- Tourist detection algorithm 1.0: Check location, if set, use. Then check location of previous tweets, if majority is outside NL, tourist. Finally, look at language of previous tweets if not english or dutch.
- Make some website that displays results of tourist detection (Ahmad)
- Event scraping (Hannes)
- Look into next step, user interest classification (Maartje)
- Idea: use dbpedia spotlight for this https://github.com/dbpedia-spotlight/dbpedia-spotlight
Example use of dbpedia spotlight from cromsom.nl shell:
curl http://localhost:2222/rest/annotate \
--data-urlencode "text=President Obama called Wednesday on Congress to extend a tax break
for students included in last year's economic stimulus package, arguing
that the policy provides more generous assistance." \
--data "confidence=0.35" \
-H "Accept: application/json"
curl -L http://dbpedia.org/resource/Economic_Stimulus_Act_of_2008 -H "Accept: application/json"
- 2016-07-13 09:00 on Skype