inst/gallery/UKTraffic/README.md

In matloff/prVis: Polynomial-Based Visualization

Data

The dataset collects the total volume of traffic on the stretch of road for the whole year in Cardiff, UK. Data can be found at here.

Graphs

prVis

All of the graphs show great separation between groups. Further, if we call the utility function addRowNums inside the prVis package, we can see some interesting trends here.

prVis  (dataset, labels = T)
addRowNums(40) # highlighting 40 rows of the dataset, and printing out
                  # the line number highlighted

# We choose 4 points lies approximately on the same line horizontally in the graph
dataset[478, ]
#busescoaches linklengthmiles frequency lab
#478       344.56             0.5    689.12   3
dataset[159, ]
#busescoaches linklengthmiles frequency lab
#159       289.85            0.62     467.5   3
dataset[1061, ]
#busescoaches linklengthmiles frequency lab
#1061       183.63            0.62  296.1774   3
dataset[631,]
#busescoaches linklengthmiles frequency lab
#631       154.79            0.81  191.0988   3

We can see that as we move along the PC1 from left to right(holding PC2 approximately unchanged), the frequency of each data point decreases. We can then infer those green points (group 3) on the far left part of the graph represents the road segments that have high frequency. One application of the insight is that those road segments may need to be improved because of the potential traffic jam caused by high volumn traffics.

tsne

umap

Citation

The data preprocessing and inspiration come from Michael Grogan: http://www.michaeljgrogan.com/uk-traffic-trends-pca/

matloff/prVis documentation built on Dec. 3, 2019, 9:27 a.m.