Temporal Graphs for Apache Spark
TemporalFrames is a Scala library for Apache Spark for representing temporal graphs. A temporal graph can be one of two data structures:
This data structure contains a column in the edge table for each time unit. Time columns must start with time_
We define a TemporalFrame as :
val temp_graph = TemporalFrame(vertices, edges)
This data structure contains a timestamp column which needs to be declared when defining a data structure. For example
val temp_graph_seq = TemporalFrameSeq(vertices, edges, 'date')
We can convert from TemporalFrameSeq to TemporalFrame using the function to_temporalframe.
val temp_graph_seq = temp_graph.to_temporalframe()
Snapshots can be extracted using the graph_snapshot function by providing the timestamp of the snapshot. If the timestamp doesn't exist, an error message will be returned. This function returns a static GraphFrame.
For example:
val snapshot = temp_graph.graph_snapshot('2018-01-05')
The timestamp function returns a dataframe with all timestamps.
val timestamps = temp_graph.timestamps()
Currently, there are 3 functions defined in the package for computing network measures:
- Burstiness
- Temporal Correlation Coefficient
- Temporal Volatility
Temporal PageRank soon to be implemented