Visualization

Visualization is handled by the companion package NarmViz.jl. It operates directly on the Rule objects returned by mine, so you can move seamlessly from mining to visual analysis.

Setup

Install NarmViz alongside NiaARM:

julia> using Pkg
julia> Pkg.add("NarmViz")

Typical workflow

1. Mine rules with NiaARM:

   rules = mine("datasets/sporty.csv", de, StoppingCriterion(maxevals=2_000); metrics=[:support, :confidence])

2. Pass the rules and original transactions to NarmViz. Consult NarmViz's documentation for available plots.

   using NarmViz, CSV, DataFrames
   dataset = Dataset("datasets/sporty.csv")
   
   visualize(
    rules[1],
    dataset,
    path="example.pdf", # path (if not specified, the plot will be displayed in the GUI)
    allfeatures=false, # visualize all features, not only antecedents and consequence
    antecedent=true, # visualize antecedent
    consequent=true, # visualize consequent
    timeseries=true, # set false for non-time series datasets
    intervalcolumn="interval", # Name of the column which denotes the interval (only for time series datasets)
    interval=3 # which interval to visualize
   )