- Occurrence can be exported with level now.
- Improved build info so the app will report ctmmweb and ctmm version more properly, make it easier to report issues.
- Updated Movebank interface to match updated Movebank API.
- Removed Unicode symbols because it can only work in certain locale/system settings and could break in user's system.
- Support IC (information criteria) in modeling step.
- Added Meta-analysis for Home range.
- Multiple updates and changes accompany with the recent ctmmweb paper. Including a new guide page for different workflow modes, many updates on help documents.
- Changed the recomanded installation method to
remotes::install_github
. - Tested with R 4.0 and the 0.5.10 version of ctmm.
- Fixed the bug with collecting diganostic info in app, which was very tricky with manipulating connections.
- Update to cooperate with recent changes in Movebank API, fixed the study table search problem.
- Updated the windows installer to R 4.0.2 and
remotes
installation method. Packages built in R 4.0.2 may not work in R 4.0.0, it's better to upgrade to R 4.0.2.
- Home range grid option will pop up automatically now.
- Windows installer is updated with R 3.6.3.
- Start page reorganized with more introduction and guides on workflow.
- Preparing CRAN release.
- Home range estimation now has the option to select either calculate in same grid or individually, because sometimes the spread out individuals can cause out of memory error. The selection must happen before estimation so the page will need user to manual trigger the calculation.
- Updated model summary code to follow changes in
ctmm
side onML
/est
changes. Older version will break with the more recentctmm
version. - Occurrence can be exported too, like home ranges.
- New windows installer will always install github master version by default. Thus the installer don't need to be updated in future.
- Added a batch script example to calculate home range in package vigenette.
- The unicode symbol in model summary table broke the generated report html. Force using utf-8 encoding now to avoid the problem.
-
Adding all
ctmm
internal datasets to app. -
Some of the datasets are anonymized, i.e. no referencee information available for location and time, only
x
,y
andt
which are relative values. A Pseudonymize process is added to simulate data with certain origin in location and time, so that the data can be analysized with app properly. -
Speed/Distance
page is added, which will estimate animals' speed and distance traveled. -
Using unicode symbols in model summary, following changes in
ctmm
-
The resolution of saved plot in
Save Progress
can be customized in sidebar. -
Fitted model
ctmm
objects will be included inSave Progress
zip, so user can load them to a R session and continue the analysis. -
Update on Movebank section:
- All attributes of data can be downloaded now, previously only several columns can be get from API. Thanks Movebank for implementing this suggestion!
- Some attributes are obsoleted and no longer used, like number of deployments.
- Better error handling when there is no data available or need license terms. Users will no longer see irrelevant warning messages.
-
Rewrite installaion instructions to make it more clear.
-
This is a official release with more tests done against
ctmm
0.5.4 CRAN release. Release version can be installed from our own repo using like this:install.packages("ctmmweb", repos = c("https://cloud.r-project.org/", "https://ctmm-initiative.github.io/ctmm_repo/"))
-
We also created a windows installer to automate the installation steps.
- Dealing with bugs, conflicts caused by
devtools::install_github
, dependency packages, alternative installation methods. - Units in tables were moved to column header so the columns can be sorted properly.
- Capture warning/error in importing data, notify user when needed.
- Take a location subset by cropping.
- Report Error information
- Calibrate data with calibration data
- Outlier page will use Error information
- The design, workflow and reactive logic have been improved
- new features added on home range weights, multiple sampling schedules and pooling variograms.
- Refit based on existing models, fine-tune model results as initial condition of next fit, showing multiple results in variograms.
- Model result page improved and condensed.
- Added much more plot control to home range, occurrence and overlap plot in a consistent way.
- export most features of app as package functions, so user can use them in their own analysis with more flexibility
- Established workflow with package functions.
- Finished vignettes for package usage.
- Changed repo name to match package name
- Moved package website to separate repo to reduce package download size and time.
- Overlap page finished.
- The app has grown to a certain degree, when there are some features can be useful independently for advanced R users. Packing app into a package make reusing these features easier, and the app distribution will also be familiar for experienced R users.
- Reference website is built
- This is a time consuming process which involved:
- changing app structure to make it work both as a function and an individual shiny app for hosting server
- untangling a lot of dependencies
- using full qualified function names everywhere
- minimizing modifications to user environment
- following other package development requirements
- building a website
- For the version after this change, you cannot launch specific version using previous method. You can install certain version of package with
devtools::install_github("ctmm-initiative/ctmmweb", ref = "v0.0.9")
ctmmweb::app()
- Interactive online map for animal locations, home ranges are added.
- For bigger dataset, heatmap is generated first since it's faster to move around in a heatmap. After you switched to the point map, the bounds of heatmap will be applied to point map so you will see roughly same area.
- Every meaningful action is logged in a html report. Report and plot pictures can be saved into a zip.
- Time consuming calculations are cached so they can rerun instantly. The cache can be save/load for next run.
- Model Selection
- Home Range
- Occurrence Distribution
- Variograms for models
- Movebank Import
- Plot individuals by selected rows or current page
- Filter outliers by distance to median center or speed
Run this version with
shiny::runGitHub('ctmm-initiative/ctmmweb', ref = "v0.0.4")
This version have much more finished features in data exploration, plots, and time subsetting.
Though the latter stages of modeling is not connected yet, because the modeling work flow may need some reconsideration to incorporate a flexible workflow into a linear page structure.
Run this version with
shiny::runGitHub('ctmm-initiative/ctmmweb', ref = "v0.0.2")
This version is a proof of concept. Every page is very simple but the app provided a complete workflow starting from import data to the final home range estimation.
Later versions will have much more sophisticated features but only work with selected pages.
To run this version you need to install shinyjs (which is not required in later versions).
install.packages("shinyjs")
shiny::runGitHub('ctmm-initiative/ctmmweb', ref = "v0.0.1")