Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
ariadne is a java based package written by Tim Evans to implement the ideas produced with Carl Knappett and Ray Rivers. It produces networks based on one of a number of models, though the primary aim is to produce networks based on the cost/benefit optimisation model we have developed. We have used this on our work on networks in the Minoan Aegean (Middle Bronze age) described in our various talks and publications (some of these are also on figshare). In principle ariadne can be used to produce network models for any problem involving two dimensional spatial networks (not necessarily geographical) and currently includes other models such as PPA (proximal point analysis) and gravity models (see interactions paper for comparison). A core part of the package is to produce visualisations (with with eps and jpg output of networks, all based on the jung network visualisations) though it can also be run without visualisations to produce statistical analysis of models. Numerical output in various formats allows visualisation and further analysis using other packages
There is some user documentation but it is not kept up to date, and the programme can be a bit delicate. So it probably needs face to face help to get it working or a good working knowledge of java (the code itself is pretty well documented using javadoc). I develop it on a Windows PC, and it should work fine on Linux. I have got it running on a Mac too despite issues there with old versions of java but it is more work to get this running.