Skip to content

Latest commit

 

History

History
74 lines (64 loc) · 3.17 KB

changes.md

File metadata and controls

74 lines (64 loc) · 3.17 KB

Review of jwh contributions and suggested changes

Data

  • I had access to several datasets
    • Master crab and env spreadsheet
    • Temporal data spreadsheet
    • jho correlation data spreadsheet
    • anova spreadsheets sent to Beth
    • crab size distribution
  • Small differences in the various sources complicated reproducing the exact results in the original draft paper. So that results are consistent, I started with the Master spreadsheet, pulled those data into R and cleaned that dataset
  • Need to:
    • move master spreadsheet to data/raw.
    • move temporal data spreadhseet to data/raw
  • Data cleaning is in R/data_prep_master.R, R/data_prep_temporal.R, and R/data_prep_size.R
    • need to update Readme with this.
  • Cleaned output csv that is basis for all analysis is data/crab_data.csv
  • burrows mp in jho correlations, and mp only in master have messed up burrow counts.

Correlations

  • Given issues above with data, correlations re-done with master data
    • slight differences in original source data for anova vs corr
  • Also, focus on signficance dropped and all correlations reported in figure.
  • correlations are calculated in R/crab_calc_correlation.R
  • correlation, p, and n are included in results/crab_correlations.csv
  • Correlation figure calced in R/corr_fig.R
  • I checked results in crab_correlations against what was reported in Table 4. All check out, except burrow_density, shear_strength, mp.
    • burrow counts messed up for burrow density in mp in the jho correlation data and in the master mp only tab. Looks good in master data and in R version of master data. Value recorded in Table 4 is incorrect. It is not 0.49, should be -0.088.
  • Need to work on figure: Color scheme, sizes aren't cutting it.
  • Filtered out pairs with less than 10 samples

Anova

  • Keep main anovas
    • was able to reproduce Table 2. with exception of carcinus. Slight diferences in the F statistic. Got this with master spreadsheet as source and also with anova spreadsheet as source. Need to update values in Table 2.
  • Drop multiple comparisons and Table 3
    • small sample size for each comparison (only 10) and within habitats are pseudo-replicates.
    • Not sure what it tells us
    • yep, lots of differences between/within marsh and habitats, but I don't find that very interesting. The big picture diff's are more interesting.
  • anovas re-run in R/anovas.R
  • anova results output in results/anova_results.csv

Temporal

  • Temporal data was reorganized by hand and is in sheet 2 in data/raw/jho_temporal_data_summarized.xls
  • Data read in an prepped in R/data_prep_temporal.R
  • Temporal figures built in R/temporal_fig.R
  • Temporal anlaysis in R/temporal_analysis.R
  • I would like to drop the water temperature plot. It is flat, currently on a plot with two y-axis, and doesn't add much to the story (IMO).
  • t-tests and values in burrow_density bar plots are not the same as what I get. Need to figure out why the difference.
  • All regressions match up
  • output in

Crab size distribution

  • Raw data in data/raw/species size histograms.xls and brrow data for histrograms.xls
  • data read in and prepped in R/data_prep_size.R
  • Size distribution figure in R/size_fig.R

Functions