| Package: |
| micromapST |
| Type: |
| Package |
| Version: |
| 1.1.1 |
| Date: |
| 2016-12-07 |
| License: |
| GPL-2 |
| LazyLoad: |
| no |
Linked micromap plots link statistical graphs to an organized set of small maps, thus adding geographical context to the graphs and data. The micromapST package has been expanded to be able to use boundary data from any collection of geographic areas through the use of border group datasets. Each border group dataset contains five R objects that define the boundaries, run parameters and name, abbreviation, and ID relationships for a geographic region. When a border group is specified in the function call, the associated dataset is loaded and the five R objects become the key data structures used by micromapST to create the linked micromaps. The five R objects are:
The currently the supported sets of border groups contain in this package are:
The default border group is USStatesBG to allow users of previous releases of micromapST to run without changes to their code. To support users of the test release of micromapSEER, a function micromapSEER is included. The function micromapSEER sets up required parameters and then calls micromapST with the border group set to USSeerBG to create the same linked micromaps as the micromapSEER package created. All of the call parameters are the same. The two packages were merged to allow micromapSEER users to benefit from new features and fixes are released under micromapST.
Additional border groups will be added over time. The user may also create their own border group for use with micromapST (see paper on creating micromapST border groups.) The entire micromap is created to fits on a single page. The page may be portrait or landscape and can range from an 8.5 x 11 up to a 11 to 17 page. Areas are grouped into panels from 3 to 5 areas each based on the sort variable, with the median-valued area set off in a separate panel in the middle of the page. If the median panel contains more than 1 area, a full link micromap panel is generated. Otherwise a single line representing the area is drawn and the median area is highlighted in the panels above and below the median. In the case of the U.S. Map and 51 states, there are 5 panels of 5 areas (states) above the median and 5 panels of 5 areas (states) below the median row.
The U.S. Seer Registry data may be groups of 9, 11, 13, 17 or 18 registries. Number of registries per panel and number of panels are dynamically setup based on the number of registries involved in the micromap. There are a variety of glyphs the caller can specify for each column of the micromap: the US map with areas colored, a list of registry names or abbreviations (the default) and one or more statistical graphics. The order of these panels is specified by the caller.
The statistical glyphs implemented in this version are plots of dots, dots with significant, dots with confidence intervals, dots and intervals based on Standard Error, horizontal bars, arrows, time series with or without confidence bands, horizontal stacked (segmented (SEGBAR), normalized (NORMBAR) or centered (CTRBAR)) bar charts, scatter plots and boxplots.
If the micromap cannot fit on one page, warnings are generated and the function is stopped. It is suggested the caller increase the size of the page (graphic space) being used to compensate.
The U.S. map of states and areas used by the USStatesBG and USSeerBG border groups are generalized boundary map, based on Mark Monmonier's visibility map. These maps are simplified to maximize the color areas shown for each state and to minimize the length of the boundary lines while still allowing identification of each area. At some future time, all border groups should have their boundaries characterized to enhance the linked micromap's readability.
One of the biggest enhancements in this version of micromapST, is support for other geographic areas. This has been added using border groups. Each border group data set contains the unique collection of information, run parameters, names, abbreviations, and boundary information for the geographic region. The package contains some pre-made border groups and the package user is encourage gather the required information and create their own border group. This is not a small task and required boundary file manipulation to reduce the complexity of the boundaries and researching to identify a suitable list of names, abbreviations and ID for each sub-area within the desired geographical area. The author is working on a guideline and a step by step procedure to help the user create their own border groups. This release includes border group to re-create the original link micromaps in earlier versions of micromapST for the 50 U. S. States and DC and micromapSEER for the 20 NCI Seer Registries. Several other test versions of border groups have been included in this release. The complete list of border groups included are:
Refer to the chapter on each border group for definitions on the Names, Abbreviations, and IDs used in the border group to link the user data to the boundary information to draw the micromap maps.
The sort order of the rows (areas) is based one of the statistical data columns as specified by the user. Correlation between multiple statistical columns can be judged visually by comparing the pattern of one column's values from top to bottom of the page with that of the sorted column. Spatial clusters of states with similar values of the sorting variable can be identified on the small maps that are linked to the graphics by color.
A area linked micromap plot is generated by 4 steps:
# load the package
library(micromapST)
# Read, create or collect your data into a data.frame. statsDFrame <- data.frame(a row per area, column per variable to be plotted, row.names set to the area names or abbreviations)
# now set up a data frame that defines the labels, # panel and page layout
panelDesc<-data.frame(...)
# specify the data source, panelDesc, sorting variable and # order, and call the stateMicromap function
micromapST(statsDFrame, panelDesc, title=c("title1","title2"), details=list(options=values))
The package contains a set of examples of how to produce linked area micromaps. The datasets used in each example are provided to help you learn how to use micromapSEER.