- Entities manipulated: airport, airline, geographical data, - The information came from geonames, ouraiports, amadeus intranet, etc .. - They are stored as: RDFs, txt, cvs, mysql databases ...
-Failed solutions: - Use all data as XML/YAML/etc and than look for the information: almost impossible to relate data this way, as it would require reading all the file a lot of times, in the worst case, making queries for relate data really slow. - Use a relational database: joining tables is SLOW. SQL is limiting. Is good to notice that, this way, is not always possible to relate all data in polynomial time.
- Why graphs: - - Why Neo4j: - scalability - graph traversal - how relationship works in neo4j (the meaning) - REST api Talk about the hierarchy to represent city -> Region -> Country -> Continent Technologies used: - Neo4j - PostGIS - Django - Lucene - Jquery - Python - Java - Javascript - HTML - XML - CSS