outviz

outviz is a R package that Greg Macfarlane and Kyle Ward built for building summary tables and output visualizations for transport planning models. The only documentation at this point are the man pages built from roxygen2 headers in the code. There are no vignettes or wiki content yet, although the former is a work in progress that will hopefully appear in this repo before 2020.

In addition to needing a better package name outviz requires plotly and chorddiag to be installed. Both packages should be installed as part of building the package locally. If the installation stalls, as it did for me, you can run install.packages("plotly") to install it from CRAN. chorddiag is an interface to the D3 library for generating chord diagrams, and must be installed from its GitHub repository:

install.packages("devtools")  # If not already installed
devtools::install_github("mattflor/chorddiag")

The package appears to export the following functions:

• chord_chart creates an HTML chord diagram from matrix information
• cross_table builds an attractive cross-tabulation
• cut_volumes cuts volumes into intervals
• link_measures_table returns a table that has yet to be investigated (and probably needs a more descriptive name)
• link_stats_table returns a table with link validation statistics
• link_targets returns the percent of links within a user-defined target range
• mdd_table returns a table with maximum desirable deviation statistics
• numerify "numerifies" a mixed vector by stripping non-numeric values from each element
• pct_error returns percent error
• pct_rmse returns the percent root mean squared error
• plot_mdd creates a maximum desirable deviation plot
• plot_validation creates a link validation plot
• plotly_mdd creates a maximum desirable deviation Plotly object
• plotly_validation creates link validation plot as a Plotly object
• plotly_tlfd plots a trip length frequency distribution comparison as an interactive figure
• prep_tlfd_data prepares model data for plotting from skim and trip matrices for input to plotly_tlfd, as well as average impedance and intrazonal percentage summaries
• rmse returns the root mean squared error

A quick look at the contents of R/data.R reveals that the following link fields are required:

• a node identifier
• b node identifier
• facility_type of the links
• facility_group is a higher-level aggregation of facility types
• area_type of the links
• area_name is a label corresponding to area_type
• volume is the assigned flow on the link
• screenline the link is associated with
• count for the calibration year
• distance in miles
• capacity of the link, presumably in same time interval as the volume
• ffspeed is the free-flow speed on the link, in miles/hour
• speed is the simulated speed from the network assignment, in miles/hour

It appears that other fields can be coded as well and will be ignored by outviz. If you want to plot chord diagrams the following fields in a tibble are required:

• origin zone, district, or region
• destination zone district, or region
• flow between origin and destintation to include in the chord diagram

As with the link records, it appears that additional fields in this tibble will be ignored by outviz.