- build first level query
- build second level query (n-ary level, it'll be generic enough) to go deeper into a city's dimensions
- build networkX digraph to traverse hierarchy
- extend the graph:
- add transactions as new nodes
- label them as transactions to differentiate them from dimensions
- connect them to their respective dimensions
- build "deeper" dimensions route in flask app
- aggregate amounts on concrete classes and propagate to abstract classes
- plot something with visjs
- only load classes nodes
- color things differently (one color for each abstract classification)
- add labels to the graph presented, trying in this order rdfs:label, skos:prefLabel, skos:altLabel and, finally, the URI if all the others fail
- join query in exploring.rq with the obeu-graph thing, this way we can build the hierarchy using the function of obeu-edges-dimensions
- check why some nodes are lost when adding labels
- build web application to navigate the levels
- add city selector (URL)
- add year selector (URL)
- improve city selector (some HTML element perhaps?)
- improve year selector
- when one of those classes is clicked it could add nodes of observations corresponding to that particular class