Node Issue

[ShobhaRajanna]

Before converting to the data object, I observed that the number of nodes in my graph was 19. However, after creating the Data object, the reported number of nodes increased to 183700, which is unexpected. I intended to have only 19 nodes in the graph. Could you please help me identify if there's an issue in my code?

[EdisonLeeeee]

That's because your node feature matrix x has a size of 183700 in the first dimension.

FYI: data.num_nodes will call x.size(0) if data does not have the property num_nodes.

[ShobhaRajanna]

After setting the 'num_nodes' parameter in the PyG Data object to 'len(nodes)', I encounter difficulties visualizing the edges when plotting the graph.

[EdisonLeeeee]

You need to first make it clear that what a node is. To be specific, you have 19 nodes but your node feature matrix x has a size of 183700 in the first dimension. They should be matched at the first dimension.

If I understand right, each node appears in df multiple times?

[ShobhaRajanna]

Yes, you're right. I initially treated each player as a separate node, but in my dataset, each player appears multiple times for every event. According to my analogy, I expected 19 nodes, representing the number of players in a game, including the ball. Do you have any suggestions for what can be done? Thank you

[EdisonLeeeee]

You can simply do x = df.groupby('player_id')[features].mean().reset_index(drop=True).values where features is the feature columns. You need also build a mapping from player_id to new node index to make sure the index in edge_index corresponds to the right player node.

[ShobhaRajanna]

Thank you

[ShobhaRajanna]

Hey! I Have one more question I've been working on a project involving a graph neural network and I've encountered a specific issue with the data representation. My dataset contains 155,640 data points. However, when I convert this dataset into a graph data object, the GNN seems to only consider unique nodes. As a result, the final graph representation is significantly reduced (Data(edge index=[2, 100], y=[1, 50, 2], node features=[10, 31, 2])). The data points represent different steps of the same entity, similar to having multiple data points for 'Kobe Bryant' with varying x and y values Do you have any suggestions on how to handle this situation and 50 and 30 are the future and history timestep.