- Sebastian Brarda [email protected]
- Felipe Ducau [email protected]
- Maria Elena Villalobos Ponte [email protected]
When providing walking directions to a destination, web mapping services usually suggest the shortest route (in terms of distance and/or time). The goal of this work is to automatically suggest alternative more enjoyable routes, which might take marginally more time but go through spots that would be more interesting to the user. In order to do that, we define a scoring function that weights each path based on social media data. Afterwards, we propose two new Graph based optimization algorithms and provide a demo UI.
- scraper: code to get data from Flickr either by making use of the API or by directly parsing HTML data
- preprocessing: methods to associate raw data to corresponding edges and to generate data for scoring
- Graph: processed data and methods for calculating interestingness, graph search algorithm implementation
- evaluation: methods to evaluate generated paths
- sightwalk: web application to showcase results (demo)
- docs: project proposal, intermediate updates, final paper and poster.
Used versions of each library is listed.
- Python v2.7.12
- numpy v1.11.2
- pandas v0.18.1
- geopandas v0.2.1
- geojson v1.3.3
- shapely v1.6b2
- networkx v1.11
- pymongo v3.3.0
Demo
- Django v1.10.3
- bower v1.8.0
- jQuery v1.12.4
- Bootstrap v3.3.7
- ion.rangeSlider v2.1.5
The demo is a django web application, in order to run it locally you should install all dependencies, go to the sightwalk folder and run:
python manage.py runserver
Then, access http://127.0.0.1:8000/paths/ it in your preferred browser.