Skip to contents

Changelog

Source: NEWS.md

EJAM v2.32.5 (July 2025)

Web app

  • Cities, Counties, States: Census units like States, Counties, and Cities/Towns/CDPs can be selected from a menu or searched by typing part of the name. Clicking “Done” will check online for the boundaries of those places, at which point the “Start Analysis” button will be enabled. Then clicking the “Start Analysis” button analyzes the sites for which bounds were found.
  • Area in square miles: The app now gets or calculates the area of each site more consistently and efficiently. (The function ejamit() has new params related to how area_sqmi() now can get square mileage info from ?blockgroupstats table without needing to download boundaries. There are new parameters called download_fips_bounds_ok, download_noncity_fips_bounds, and includewater. The new params are also driven by two new defaults in global_defaults_shiny.R The old parameter default_download_fips_bounds_to_calc_areas is no longer a param in ejamit()).
  • County population counts: Fixed county population counts obtained from and shown in some maps (via fixes in a function used by shapes_from_fips() so, e.g., if using mapfast(), mapfast(shapes_from_fips(testinput_fips_counties)) now shows the right numbers).
  • Summary Indexes (aka EJ Indexes) had some incorrect numbers, so this release has replaced ?bgej dataset with correct numbers. (Correct numbers were drawn from the internet archive version that was a copy of the datasets EPA had posted August 2024).
  • Sort order of FIPS Census units: Sort order of output FIPS codes and polygons should now always be the same as the order of the inputs (sorted like they were in an uploaded shapefile, uploaded FIPS, or FIPS selected from the dropdown list).
  • Medians in barplots: DRAFT feature/ work in progress – interactive barplots of indicators will be able to show median not just mean (via the ejam2barplot_indicators() function).

RStudio users only

Documentation updates

Functions added or improved

  • Mix of fips types allowed:
    • shapes_from_fips() now accepts a mix of city and noncity fips (state, county, tract, blockgroup), so you can get a shapefile where some polygons are cities and others are counties, etc. Previously that was not possible and caused an error. See parameter allow_multiple_fips_types in shapes_from_fips().
    • getblocksnearby_from_fips() now accepts a mix of city and noncity fips (state, county, tract, blockgroup), so you can get a shapefile where some polygons are cities and others are counties, etc. Previously that was not possible and caused an error.
  • fips2name() now also provides text name for a tract
  • mapfast() for a single point now zooms out enough to see the whole radius (e.g., mapfast(testpoints_10[1,], radius = 10))
  • mapfastej_counties() has improved color-coded maps of counties.
  • convert_units() now can recognize more abbreviations like “mi^2” via updated fixnames_aliases(), and got some bug fixes.
  • fips_bg_from_latlon() drafted as unexported function that identifies which blockgroup each point is inside.

Functions fixed or modified

  • ejamit() and shapes_from_fips() (and related helper functions) have more consistent, useful outputs:
    • Sorting: The outputs now consistently preserve sort order of the input (points, fips, or polygons). This had not been the case for shapes_from_fips() outputs, and the table results_bysite from ejamit() or doaggregate() was preserving sort order only for the latlon case but not necessarily the fips or shapes cases.
    • Invalid sites: The outputs of shapes_from_fips() (and related helper functions) will have a row for each valid or invalid input site (it will no longer omit output rows for invalid fips and when boundaries could not be obtained for valid fips) – The number of rows in a shapefile output will be the same as then length of the input fips vector. The output table results_bysite from ejamit() also has a row for each valid or invalid input site. That table in the output of doaggregate() in contrast does not have a row for any site lacking blocks, since the input is from getblock_xyz functions (getblocksnearby(), getblocksnearby_from_fips(), get_blockpoints_in_shape()), which don’t provide those sites.
    • Columns from shapes_from_fips() and related helpers: The output columns are ordered in a more useful way and are more consistent across functions. The output also consistently tries to add population, area in square miles, name of census unit, state abbreviation, etc., via new helpers like shapefile_addcols()
  • getblocksnearby() and related functions (getblocksnearby_from_fips(), get_blockpoints_in_shape(), etc.) also have more consistent outputs:
    • Unique ID in FIPS case: The ejam_uniq_id column in the outputs of these functions will be based on 1 through the number of sites in the inputs (with multiple rows per site as needed to include all the blocks). Previously, FIPS codes had been used as the ejam_uniq_id sometimes (and still are in the outputs of functions like ejamit() where the output has a table with one row per site).
    • Sorting: The output sites are now sorted like the input sites (points, fips, or polygons), while there are still usually many rows (blocks) per site. It had been sorted primarily by blockid, previously.
    • Invalid sites: The outputs of all the getblock… functions will be consistent – They all provide a sites2blocks data.table output (like ?testoutput_getblocksnearby_10pts_1miles) that does not include any sites that have zero blocks. The ejam_uniq_id will still correspond to the input vector, so if an invalid and valid site were input in that order, 2 would be the only ejam_uniq_id in the sites2blocks table. The FIPS-based functions, though, like getblocksnearby_from_fips(), when returning a spatial data.frame, will include all the sites in the output, even if they have no blocks, so that the number of rows in the output shapefile will match the number of sites in the input fips vector.
  • shapes_from_fips() (and related) have new year parameter, passed to [tigris::places()], defaulting to the 2024 boundaries polygons of cities/towns.
  • testoutput_xyz .xlsx and .html files and dataset R objects like ?testoutput_ejamit_100pts_1miles have been updated to reflect the new ?bgej dataset, typo fixes, and other edits.
  • Some testinput objects like testinput_fips_counties are now vectors per is.vector(), and no longer have metadata stored as attributes like date_saved_in_package, etc. Adding that info via metadata_add() was making is.vector() FALSE and interfered with some functions that expect the input to be a vector, like shapes_from_fips(). Also, testinput_xtrac was removed.
  • doaggregate() and ejamit() now report 0 for results_bysite$blockcount_near_site and results_bysite$bgcount_near_site if there are none, and total counts are correct.
  • getblocksnearby() based on getblocksnearbyviaQuadTree() will no longer include, in its output, the lat lon columns from the input table of sitepoints. That was unintentional and potentially confusing and wasted space.
  • plotblocksnearby() rewritten to fix/improve map popups, etc., and a parameter was dropped

Package development/ technical

  • Many unit tests added, especially for doaggregate() and getblocksnearby_from_fips() and related.
  • test_ejam() is what used to be called test_interactively() – it was improved and renamed and moved to the R folder as an unexported internal function loaded as part of the package. Also, a new parameter y_skipbasic is used instead of y_basic.
  • test_coverage_check() utility was improved (but somewhat work in progress), just as a way to for package maintainers/contributors to look at which functions might need unit tests written.
  • Utility functions related to package development were renamed, e.g., in utils_PACKAGE_dev.R
  • linesofcode2() utility was improved, just as a way for package maintainers/contributors to look at which files have most of the lines of code, are mostly comments, etc.
  • table_xls_format_api() is what used to be called table_xls_formatting_api() (but is not used unless the ejscreenapi module or server is working).
  • fixed inconsistent use of parameter in_shiny versus inshiny, to always call it in_shiny
  • removed functions and text related to pins board (obsolete)
  • renamed map_headernames spreadsheet file to reflect a new version (EJAM/data-raw/map_headernames_2.32.5.xlsx), made edits/fixes (spelling of CEJST, e.g.), and updated the data object ?map_headernames.
  • rebuilt favicons per updates in {pkgdown}
  • Edited DESCRIPTION file to specify minimum versions for most packages in Imports, and a newer version of R. Almost all of these just refer to the latest version on CRAN as of this release, even though several were not strictly necessary for the functions to work correctly.

EJAM v2.32.4 (June 2025)

Note the URLs, emails, and notes about repository locations/owners were edited to reflect this forked non-EPA version of the EJAM package being located at ejanalysis/EJAM, so the package called the v2.32.4 release on ejanalysis/EJAM is slightly different than the version called the v2.32.4 release that was released on USEPA/EJAM-open.

Web app

  • Fixed logo in “About” tab, app header, and report header, in app_ui, generate_html_header(), global_defaults_xyz, etc., and updated testoutput files related to ejam2report() and ejam2excel()
  • corrected spelling in app and documentation
  • added better examples of params one can pass via run_app()

RStudio users only

Technical / internal changes:

  • enabled testing of web app functionality from the test_interactively() utility (which has more recently been renamed test_ejam() and put in R folder as an unexported internal function loaded as part of the package) or via test_local(), etc., not just from a github action. (See /tests/setup.R which now has a copy of what is also in app-functionality.R)
  • drafted revisions to ui and server to try to allow for more run_app() params or advanced tab or global_defaults_xyz to alter default method of upload vs dropdown, e.g., output ss_choose_method_ui, default_ss_choose_method, default_upload_dropdown. This included revising server and ui to use just EJAM:::global_or_param() not golem::get_golem_options(), so that non-shiny global defaults can work (e.g., logo path as global_defaults_package$.community_report_logo_path) even outside shiny when global_defaults_package has happened via onattach but global_defaults_shiny etc. has not happened.
  • changed .onAttach() to do source(global_defaults_package) with local = FALSE not TRUE, but this might need to be revisited – note both local = F and local = T are used in .onAttach() versus get_global_defaults_or_user_options()
  • in server, ejam2excel() now figures out value of radius_or_buffer_description, ejam2excel() gets new parameters table_xls_from_ejam() uses improved buffer_desc_from_sitetype() and now uses ejam2report() to add a report in one tab.
  • reorganized server code by moving v1_demog_table() and v1_envt_table to long report section of server file
  • cleaned up server code (eg, remove obsolete input$disconnect, remove obsolete community_download() and report_community_download(), and remove repetitive ejam2repor(), remove old EJScreen Batch Tool tab, used session = session as param in server calls to updateXYZINPUT, etc.)
  • allow shiny.testmode to be TRUE even if not set in options
  • used silent=TRUE in more cases of try()
  • added validate(“problem with map_shapes_leaflet() function”)
  • added validate(need(data_processed(), ‘Please run an analysis to see results.’))

EJAM v2.32.3 (May 2025)

  • Added a long list of additional indicators in the summary report (in a subtable) and in outputs of ejamit(), etc. New indicators include counts of features (Superfund sites, schools, etc.), asthma and cancer rates, overlaps with certain types of areas (Tribal, C JEST disadv., air nonattainment areas, etc.), flood risk, % with health insurance, more age groups (% under 18), and numerous other indicators. You can see the expanded report via ejam2report() or at system.file("testdata/examples_of_output/testoutput_ejam2report_100pts_1miles.html", package = "EJAM")
  • Area in square miles (area_sqmi column) added to results, with calculation of size of each location (polygon or FIPS unit or circle around a point)
  • More/better info on number of sites or site ID and lat/lon, now in header
  • Enabled customization of summary table (for R users) to show fewer or new additional indicators (as long as they are in the outputs of doaggregate() and ejamit() or at least are in the inputs to ejam2report() etc.). This is done via the extratable_list_of_sections parameter in ejam2report(), in build_community_report(), in the community_report_template.Rmd, and in global parameter default_extratable_list_of_sections. It may later be enabled as modifiable in the advanced tab.
  • Easier to set which logo to show on summary report (EPA or EJAM or other logo), in global settings

Other web app improvements

  • More types of shapefiles can be uploaded in the web app – json, geojson, kml, zip (of gdb or other), and shp.
  • Census units like States, Counties, and Cities/Towns/CDPs can now be selected from a menu or searched by typing part of the name, in a shiny module called fipspicker, and the feature is enabled/disabled via global settings use_fipspicker and default_choices_for_type_of_site_category. It works but current does not check or alert users if boundaries are not available, until after the Start Analysis button is clicked.
  • Simpler UI for “More info” button about file types and formats allowed in upload.
  • Preview maps can show FIPS now, along with shapefile polygons, or points
  • ejam2report() and ejam2map() and mapfast() now better able to create maps of polygon data, FIPS, one site vs all sites, etc.
  • progress bar added for doaggregate() in cases of fips and latlon
  • Clarified/explained 2025 status of API and urls in CONTRIBUTING and README, etc.
  • Extensive additions of and improvements in articles/vignettes, including documentation of how to maintain repo, package, and datasets. Articles/vignettes avoid hardcoded repo urls, and use relative links within pkgdown site… unexported helper function EJAM:::repo_from_desc() added, avoids hardcoded repo url; download_latest_arrow_data avoids hardcoded repo url; links to testdata files on webapp UI avoid hardcoded repo url; simpler What is EJAM doc.
  • ejamit() in interactive mode (RStudio) now lets you select any type of file to upload if no sites specified by parameters
  • Many options or starting values or settings for the shiny app (and in general) can now be set as parameters passed to the run_app() function, which overrides the defaults. extensive changes to global defaults vs user parameters allowed: replaced global.R; files renamed, put in 1 folder, etc. System for using user parameters passed to run_app(), global defaults otherwise, many can be changed in advanced tab; some may be bookmarkable. The default values are now set for the shiny app and in general in files called global_defaults_package.R, global_defaults_shiny_public.R, and global_defaults_shiny.R (rather than in the old files global.R or manage-public-private.R).
  • acs_bybg() examples added, on how to obtain and analyze new/custom indicators from the American Community Survey (ACS) data
  • testdata() function improved, showing you examples of files that be used as inputs to ejamit(). testdata() files and data objects cleaned up/renamed consistently and new ones added for fips types, naics, sic, mact, etc.
  • refactored names of plot functions made more consistent to use “plot” singular and “ratios” plural, as in ejam2boxplot_ratios(), boxplot_ratios(), etc.
  • documentation fixed in some functions (e.g., ejam2map())
  • large datasets managed via dataload_dynamic(), download_latest_arrow_data() and other new arrow-related functions and no longer on pins board or aws at all. arrow datasets faster format used most places, other changes to handling downloads etc.
  • shape_from_fips() checks if census API key available and tidycensus pkg now imported, uses alt method (arcgis services) to get boundaries if necessary.
  • Continued towards refactoring/consolidating code in server vs in functions, related to creating summary report as HTML vs for download from shiny app vs from ejam2report(), in helper functions such as build_community_report(), report_residents_within_xyz(), renamed generate_demog_header to generate_env_demog_header, etc.
  • server uses ejamit() for SHP and latlon, and cleanup
  • server uses ejam2excel() now (which then relies on table_xls_format())
  • server uses ejam2report() now, not obsolete report_community_download() etc.
  • server uses shapefile_from_any() now
  • server: removed use of data_summarized reactive everywhere, use data_processed$…
  • 2 new params doaggregate() has, to ejamit(), for calctype_maxbg and minbg
  • bug fixes such as in ejamit() for wtdmeancols param, ejamit_compare_distances(), shapes_from_fips(), plot_ridgeline_ratios(), map_google(), in mapfast() for tracts vs blockgroups, many others
  • unit tests added and others updated/fixed
  • misc helpers/utility added/updated/documented
  • renamed .xlsx file of map_headernames info to reflect a new version and made edits/fixes
  • reposissues() and repoissues2() help record snapshot of gh issues
  • DESCRIPTION file now has new field ejam_data_repo
  • updated workflow action to use latest version of github-pages-deploy-action

EJAM v2.32.2 (February 2025)

  • Revised all language based on executive orders, to refer to environmental and residential population data analysis, rather than EJ/EJScreen/etc.
  • Revised web links based on EJScreen website being offline
  • Some edits made considering github repositories and gh pages may change location or go offline
  • Updated FRS datasets, pulled on 2/12/25
  • Remove screenshots from user guide document

EJAM v2.32.1-EJAM (February 2025)

Bug Fixes

  • Fixed metadata warning shown during loading of arrow datasets
  • Fixed typos in languages spoken indicators labels
  • Improved labeling and legibility of barplot of ratios used in reports and downloads
  • Fixed caps to # of points selected, analyzed

Enhancements

  • Expanded tables of indicators shown in community report
  • Languages spoken at home, health, community, age
  • Added ratio columns to community report as advanced setting and heatmap highlighting optional
  • Incorporated shinytest2 tests for app-based functionality testing
  • Implemented mapping for points in ejam2excel()

Experimental enhancements

  • Added draft plumber API for ejam2excel()
  • Added widget to advanced settings
  • proxistat() helps build proximity indicator
  • Zipcodes vignette

Other

  • Refactored community report functions, app_server.R script

EJAM v2.32-EJAM (January 2025)

New Features + Improvements

  • Enabled automatic download of latest arrow data from ejamdata repo
  • Incorporated public-internal toggles to hide specific UI elements not yet applicable to the public version of EJAM
  • Made improvements to maps of polygons
  • Added shapefile upload instructions

Bug Fixes and Enhancements

  • Added leaflet.extras2 dependency to Imports, instead of Suggests, which is necessary for new installations

EJAM v2.32.0