- Minor update to globals to support dplyr update.
- Refactored codebase to use latest dplyr syntax
- Added support for logical outcome and categorical predictor variables in
create_IV() - Added detection of text missing values in
validation_report() - Enhanced flexibility in display control for
create_dt() - Improved test coverage
- Improved missing value detection for
validation_report(),hrvar_count_all()(#35) - Added support for logical outcome variables for
create_IV()(#240)
- Minor bug fixes and aesthetic improvements.
Updates to network_p2p(), with breaking changes to the argument, addressing issue #225:
network_p2p()now has acentralityargument for computing and visualizing centrality as node sizes in the network plot- the
algorithmargument is renamed tolayoutfor better intuitiveness - Improved consistency and intuitiveness of the API, with
styleargument now controlling the network plotting mechanism andreturnargument controlling whether plots are generated interactively or saved as PDF - Added a large number of community detection algorithms from
igraph
network_summary() adds the ability to return pagerank.
This version also includes some minor dependency maintenance done to create_dt().
- fixed issue due to a 'tidyr' update (see #233).
- Updated and improved output and algorithm for
workpatterns_classify() - Additional visual options for
workpatterns_classify()andflex_index() - Added the
external_*family of functions to visualize the metricCollaboration_hours_external - Added experimental function
identify_datefreq()for detecting date frequency (#131)
- Renamed 'Workplace Analytics' to 'Viva Insights'
- Improved features for multiple functions, e.g. centrality calculations for
network_p2p() - Improved compatibility of functions with new queries
- Introduced new functions, e.g.
create_density(),keymetrics_scan_asis(),network_summary() - Removed 'portes' dependency, which was removed from CRAN
- Bug fixes
- Made
meeting_skim()dynamic - the function will still run even if not all variables are present. (#198) - Enable text mining functions to accept
stopwordsargument as a string vector, rather than a data frame (#182) - Better handling of errors due to variable name inconsistency when loading in a Ways of Working Assessment query as opposed to a Standard Person Query (#201)
- Added
create_inc()for creating incidence analysis - Replaced outdated uses of
mutate_all()in functions (#195)
- Added new signal options for
flex_index()(#183, #185) - Minor feature improvements (#186)
- Minor bugs fixed (#181, #184)
- Updated
em_datawith new signals (#191)
- Fixed several minor bugs reported on GitHub (#177, #178, #180, #181)
- Refurbished
meeting_tm_report()(#173)
- Improved visuals on
workpatterns_rank()andidentify_shifts_wp()(#160) - Added ability to show percentages on
identify_shifts_wp()(#160)
New functions:
generate_report2()(#141)wellbeing_report()(#141)create_tracking()(#141)anonymize()(#162)jitter_metrics()(#162)
Other changes:
- Improved error handling on
create_IV()(#157) - Bug fixes to plotting mechanism for
network_p2p()and added ability to select custom weights (#149)
- Improved aesthetics on plots
- New visuals, with abbreviated labels added to archetypes generated by
workpatterns_classify()(#130, #132)
New functions:
hrvar_trend()(#132)one2one_freq()(#132)
- Patch update
- New CRAN submission addressing some feedback on Rd files (successful)
- Patch update
New functions:
one2one_freq()(#95)
Significant changes to existing functions:
create_rank()is now able to callcreate_rank_combine()to compute averages for pairwise combinations. Themodeandplot_modearguments are also made available to the key metric wrapper functions ofcreate_rank().network_g2g()now has the option to specify algorithm and node colour.
v1.4.0 is the version where the working patterns family of functions are released open source as part of wpa (#55). These functions include:
workpatterns_classify()flex_index()workpatterns_report()plot_flex_index()
In addition to the above, there are also some new functions, bug fixes, documentation, and performance improvements.
New functions:
import_to_fst()(#86)create_hist()(#80)identify_shifts_wp()(#66)
Significant changes to existing functions:
collaboration_report(): visuals and structure are improved (#69)import_wpa()can now specify encoding when reading in data, which now defaults toUTF-8(#76). This resolves a previous bug for reading in double-byte characters.network_g2g(): now auto-detects columns for collaborators and time investors. Also, if the value'Collaborators Within Group'does not exist in the data, a warning message is returned.
Bug fixes:
collab_area()now has relaxed variable name checks which makes it less prone to fail when not all exact collaboration metrics are available or when cases do not match (#81).- Resolved an issue of overly strict checks on
flag_outlooktime(). Now it is possible to pass both7:30and07:30as inputs.
Documentation changes have been implemented across the board to comply with CRAN and to improve the user experience (#31)
New functions, bug fixes, and performance improvements.
Significant changes to existing functions:
- New plot visual is available for
keymetrics_scan() combine_signals()can now dynamically accept any metrics available in the Hourly Collaboration query.pairwise_count()now uses a data.table implementation, instead of dependent on widyr.
New functions:
network_p2p()network_leiden()network_louvain()network_describe()create_sankey()totals_col()
Some package dependencies have been removed (see #36):
- network
- GGally
- widyr
This is the first version of the wpa package to be released open-source on GitHub. If you have been using a previous developmental version, the main difference is that this release omits the more experimental working patterns family of functions. The experimental functions are currently available upon request via mac@microsoft.com.