1

I have a CSV file with bus route information that looks like this. I am having trouble creating nodes and path relationships in Neo4j with it in this format.

I would like to have nodes for the stops and routes, and routes between them using the sequence and route detail id to show the direction of the routes.

RouteName route_detail_id Stop Sequence Arrives Departs
Bus1 50701 Cherry 1 9:00
Bus1 50802 Market 2 9:30 10:00
Bus1 59003 Raleigh 3 10:30 10:50
Bus1 59004 Stuart 4 11:05 11:30
Bus1 58006 Possum 5 12:30
Bus2 67003 Cherry 1 11:00
Bus2 67004 Market 2 11:30 12:00
Bus2 67009 Raleigh 3 12:30 12:50
Bus2 67010 Stuart 4 13:05 13:30
Bus2 67011 Possum 5 14:30
Bus3 89004 Highland 1 9:00
Bus3 88005 McKinley 2 9:30 10:00
Bus3 67098 Jersey 3 10:30 10:50
Bus3 4500 Ridgewood 4 11:05 11:30
Bus3 67890 Osprey 5 12:30

route_detail_id is the unique identifier for that particular stop on that particular route.

I would like to be able to use the times for shortest path queries in the future, but right now would just like to be able to create a structure and visualize in neo4j.

Eventually it will be used to create connecting routes, and shortest path searching, but right now I am just stumbling over even converting information in this format to Neo4j.

Carmanders
  • 11
  • 1

1 Answers1

0

I would start by converting the format into a list of nodes connected by arcs, such as:

  1. Cherry -- Bus1, 50701, n/a, 9:00 --> Market
  2. Market -- Bus1, 50802, 9:30, 10:00 --> Raleigh
  3. ...
  4. Cherry -- Bus2, 67003, n/a, 11:00 --> Market
  5. ..

This seems to me to be a more natural way of representing the data, as you have stops (nodes), which are connected by bus routes (directed arcs, with route details).

You can then query the database by looking for links between the nodes. You can convert also convert the arrival/departure times into the duration of the journey between two nodes if you want to find a shortest path.

Oliver Mason
  • 2,240
  • 2
  • 15
  • 23
  • Yea! That’s what I want to do, I don’t like it in this format but I was hoping I could just pull it into Neo4j in that format and can’t quite figure that out! Maybe I’ll just throw it in pandas or something and convert. – Carmanders Mar 18 '22 at 16:58
  • I’ve tried a couple different methods in neo4j with apoc pairs etc but they don’t like the null values even though no null should be carried over so I’m doing something wrong – Carmanders Mar 18 '22 at 17:00