search

SiRumCz / CSC501

CSC501 assignments
0 stars 1 forks source link

Node link diagram #79

Open jonhealy1 opened 4 years ago

jonhealy1 commented 4 years ago

For this diagram we need weights. In our data right now, some nodes have like a 100 connections between them and we have to somehow convert the number of connections into a weight I think. Is there something else that we were thinking of doing for weights?

We could have two connections from node to node - one for positive and the other for negative sentiment. We also need to constrain the number of nodes we use - maybe just use the top 20 or 30 nodes.

soroushysfi commented 4 years ago

For this diagram we need weights. In our data right now, some nodes have like a 100 connections between them and we have to somehow convert the number of connections into a weight I think. Is there something else that we were thinking of doing for weights?

We could have two connections from node to node - one for positive and the other for negative sentiment. We also need to constrain the number of nodes we use - maybe just use the top 20 or 30 nodes.

Yeah with this diagram we can't show more than 40-50 nodes(it also depends on the number of edges if the edges aren't that much we can depict graph with up to 100 nodes). The properties section of edges provides some weights that we can filter out edges(like average word count). If we turn connections into weights we would probably loose some source or targets, right?

jonhealy1 commented 4 years ago

That's true we could use other properties instead of just link sentiment.