Warning Signals - Manuscript touch-ups

[toc]

Refining examples

  • Updating finished runs to full formats

  • Condensing/organizing code examples under the demos to follow the structure of the paper.

  • Currently plotting code for manuscript loads a menagerie of different run results from different files.  Would be better to consolidate by dataset.

Proposed figure changes

  • Consider multiple empirical examples in main text

  • Consider removing skew/kurtosis examples to the supplement?

  • Consider other empirical examples

  • Additional examples to the supplementary information, at least of a case where likelihood also has inadequate power

  • Consider a low-sample size version of ibm simulation

  • Consider replacing the linearized simulation with the ibm simulation

Figure formatting guidelines (submission stage):

  • Preferred format is jpeg, 150 dpi

  • figures rendered in the size they appear in the manuscript (single column is 89mm, double column width is 183 mm)

  • follow naming convention Boettiger_fig1.jpg, etc

  • … Getting images adjusted for small size – that took a while, seegit commit.

  • shrink ticks, move text, set axis intervals.

  • All figures set to single column width, decide if any should be double-column?  (doesn’t seem typical).

Revisions and Edits

  • Incorporated Marcel’s edits.

  • Awaiting Marissa’s edits.

  • Just sent to Alex for edits this morning, back this afternoon.

  • Incorporating Alex’s edits now.

  • Reviewed** cover letter **with Alan, adding details about figures, word counts, etc. awaiting comments on my revisions.

Further Examples

  • drake reanalyze figure3 failed (nice 16 on zero) due to labeling error in plotting function (after all computations finished but no data save – highlighting the need for checkpointing my code!)

  • Rerunning at nice 19 on zero 16 reps, Completed

  • and at nice 18 the full 2000 replicates on zero.

  • ibm_likelihood.R terminated(?) Running at 80 replicates at nice 17 on zero