← Back to all sparks
N

ncmR

INFRA · APIS
Velocity0.0

Fit Neutral Community Model to Microbiome or Ecological Data

Three weeks from first CRAN submission to a point-and-click neutral community model

microbiomeecologyneutral-community-modelshinyearly-stage
Current state
ncmR fits the neutral community model to microbiome and ecological abundance data. The whole feed is three weeks long: fit_ncm() and summary methods at 0.1.0 on 1 April, a scatter plot with fitted curve and confidence interval at 0.2.0, then a Shiny application wrapping both at 0.3.0, then a row-name bugfix. It is a new package moving fast in its first month.
Where it's heading
The direction is toward users who do not write R. Version 0.2.0 added plotting, 0.3.0 wrapped fitting and plotting in Shiny modules, and the same release deleted the Unicode plotting helpers introduced one version earlier because the dependency they existed for was removed. That willingness to throw away a week-old API suggests the surface is still being negotiated rather than settled.
Prediction
With a fitting module and a plotting module in the app, the remaining gap is getting results back out — export of fitted parameters or figures from the Shiny session. The bugfix at 0.3.1 was in file upload, which is where a GUI's problems usually start.

Recent moves

  1. 4mo ago

    Row names preserved during file upload conversion

    A first bug report against the two-day-old Shiny app: converting uploaded data frame columns to numeric dropped row names. Exactly the class of defect a new GUI surfaces.

    View source ↗
  2. 4mo ago

    A Shiny app for fitting and plotting neutral community models

    ⚡ SPARK

    Nineteen days after its first CRAN submission, ncmR gains an interface. The app ships Fit NCM and Plotting modules over the functions from 0.1.0 and 0.2.0, and the Unicode helpers added in 0.2.0 are removed in the same release now that plotting no longer depends on them.

    View source ↗
  3. 4mo ago

    Scatter plot with fitted curve and confidence interval

    The canonical NCM figure — occurrence frequency against mean relative abundance with the fitted curve and its confidence band — becomes one function call. The Unicode display helpers shipped alongside it were removed at 0.3.0.

    View source ↗
  4. 4mo ago

    First CRAN submission: neutral community model fitting in R

    ⚡ SPARK

    The package's first appearance: fit_ncm(), summary() methods for NCM and NCM_group objects, and an example dataset. Everything in the two weeks that follow is built on this surface.

    View source ↗