Graph Dataset: PostgreSQL or Neo4j?

Question

I am pretty much a beginner, and for a personal project have decided to map a relatively straightforward dataset describing some internet connections. I know a bit of SQL and worked previously with Postgres. But I am curios also about Graph databases such as Neo4j.

The dataset is composed by 3 tables:
A - (nodes)
nodeID: i1 i2 ... in
N33 91.6.6 44.55 99.32.1

B - (links)
linkID N1:i1 N2:i2 ... Nm:im
link99 N33:5.32.4 N14:11.49.66

C - (node attributes)
nodeID country region city latitude longitude
N54: CH WF Lausanne 19.3267 34.8167

As an additional information I might add that these data have a spatial component, and I would like to come up with a visualization that takes that into account (ie. map the nodes/links according to their lat/lon locations).
This is a crucial point for me.
I know Gephi has a geo-plugin, but I also know that it requires a precise data structure, which at the moment differs from mine, ie:

links:
source | target
1 | 2
1 | 43
1 | 27
nodes:
nodeId | attribute1 | attribute2 ...
1 Lausanne CH

My question in broad terms is thus: what should be my best options to import and query/visualize the data?

• should I stick to PostgreSQL?
  • But then, how can I import and manage the data from table A and B (which have a varying number of columns)?
    Approach it like a many-to-many relationship issue?
    Should I import the tables with the largest amount of columns possible and treat the rest as null values?
• should I re-shape the data beforehand? thus making them more similar to Gephi data structure (with the source node spanning multiple rows and only another column: the target column)
• should I try the Neo4j way? I am willing to try it and from what I heard it's pretty neat, though I am a beginner and not sure if it's suitable for my skills set. Also, isn't it maybe overkill for this dataset?

Thanks everyone in advance for your answers, and sorry for the long post.

Answer 1

I think a lot of it depends on how you want to query the data. I'm not familiar with the capabilities of Gephi with respect to spatial, but there is a plugin for Neo4j directly which does spatial querying:

https://github.com/neo4j-contrib/spatial

Generally, though, that part is sort of perpendicular to the traversal query power of Neo4j. With spatial queries you're generally narrowing a set of nodes down by an area. You might then optionally browse across the links between nodes.