jesstytam/honours

💣 Task manager

⛏️ Data collection

• Body mass (5400 spp)
  • EltonTraits 1.0 (will only use this for now)
    • collected
    • cleaned
      - [ ] ADW
      - [x] scraped
      - [ ] cleaned
      - [ ] AnAge
      - [x] collected
      - [ ] cleaned
      - [ ] PanTheria
      - [x] collected
      - [ ] cleaned
      - [ ] Smith et al.
      - [x] collected
      - [ ] cleaned
• GBIF geographical location (4721 spp)
• Phylogeny (6952 spp)
• Human use (1472 spp)
  • collected from IUCN
  • try pivot_wider()
• IUCN status (5934 spp)
  • collected
  • cleaned

🧮 Pre-analysis

• Combine datasets (8308 spp)
• More cleaning up
  • remove species with only genus or species name (8305 spp)
• Add classification levels (taxize)
  • will have to come back to do this after itis is working again
• Synonym matching (rotl)
  • create a species list in long form with the ID
• h-index (specieshindex)
  • put quotation marks around synonyms
  • fix weird NA dataframe (putting synonyms around the synonyms seems to have fixed the problem)
  • divide species list into 2 to fix timeout issue
  • run the 2nd list next week since I’ve reached the limit of 20,000 requests this week
• Missing data (mice)
• Google Trends (gtrendsR)

🎨 Making graphs

• h vs mass
• h vs phylogeny 📍
  • something wrong with the new artificial tree, need to find out why
• h vs location
• h-index map
• h vs human use
• h vs domestication
• h vs iucn
• h with vs without conservation keyword
  • check for patterns
• Google Trends sum
• For all plots
  • change font
  • change h-index axis scale to true values
  • add p-values / other stats on the plots

🚀 Statistical analysis 💎📍

• correlation matrix of complete entries
• Phylogenetic signals (phytools)
• MCMCglmm
  • impute 10 datasets
  • trim all trees + pick 50 randomly
  • write nested loop + TESTING
  • katana
  • extract data from models 📍
  • Pagel’s λ (calculated from MCMCglmm results) :round_pushpin:

🧩 Other materials 💎📍

• data summary table for thesis
• supplementary material

 

Legend
What I’m working on now 📍
Priority 💎

honours's Issues

iucn status

check for missing data

check if mammals with higher h-index have less missing data
convert data to binary (1 for data present, 0 for data missing)

check this out - altmetrics

@jessicatytam - have a look

https://conservationbytes.com/2020/11/17/collect-and-analyse-your-altmetric-data/

finalize species list for study

possible workflow:

• get vector of names from ROTL
• check overlap with body size data see #3
• filter GBIF to this species list

ultrametric trees

@wcornwell how do we make a non-ultrametric tree ultrametric?

Get a phylogenetic tree(s)

@jessicatytam will get @mlagisz's help to get a tree for mammals

plotting relationships!

ggplot - log10(h) vs log(BodyMass) etc. Also, you can explore relationships among all variables
ggplot - log(h) vs IUCN status and human use

parsing animaldiversity.ummz.umich.edu

@jessicatytam has got all the data but it's coming in like this:

Cryptotis squamipes\n      \n      \n      \n      \n      \n    \n        Cryptotis thomasi\n      \n      \n      \n      \n      \n    \n        Ctenodactylus gundi\n      \n      185\n      175\n      195\n      \n    \n        Ctenodactylus vali\n      \n      \n      \n      \n      \n    \n        Ctenomys argentinus\n      \n      \n      \n      \n      \n    \n        Ctenomys australis\n      \n      \n      \n      \n      \n    \n        Ctenomys azarae\n      \n      \n      \n      \n      \n    \n        Ctenomys boliviensis\n      \n      \n      \n      \n      \n    \n        Ctenomys boliviensis nattereri\n      \n      \n      \n      \n      \n    \n        Ctenomys bonettoi\n      \n      \n      \n      \n      \n    \n        Ctenomys brasiliensis\n      \n      \n      \n      \n      \n    \n        Ctenomys colburni\n      \n      \n      \n      \n      \n    \n        Ctenomys conoveri\n      \n      900\n      \n      \n      \n    \n        Ctenomys dorsalis\n      \n      \n      \n      \n      \n    \n        

so we need some clever regex help from @mlagisz

which model to use?

need to use a model with log values
model with real or log values?

h-index map

getting hits from the internet

https://cran.r-project.org/web/packages/gtrendsR/gtrendsR.pdf

learn regular expression

https://www.youtube.com/watch?v=q8SzNKib5-4

finalize selection of body size data

discussion of different data sources: see also #2

latitude

match synonyms in trees

match synonym in the trees
5,000+ matches would be ideal

making a map of where all the species are in space

Have to first make a calculation of species of species centroid:

https://rmacroecology.netlify.app/2018/02/06/species-level-macroecological-analysis/

common name matching

go to iNaturalist

correlation matrix

common name matching encoding error

resolve error "No encoding supplied: defaulting to UTF-8." or rinat common name matching

MCMCglmm

since i didn't match the synonyms properly before i did it again and got a pretty similar list of mammals.
i imputed data from this new list, trimmed the tree, and run the model again, but now body mass is not significant (everything else is the same), is this bad?

post.mean l-95% CI u-95% CI eff.samp pMCMC
logmass 0.03489 -0.01843 0.08686 1117 0.204

produce summary graphics of species list

this is for after #4

fill in missing iucn status

fill in manually

Reading list

@jessicatytam - please check this out

phyloregion: R package for biogeographical regionalization and macroecology - Daru - - Methods in Ecology and Evolution - Wiley Online Library

https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.13478#.X2oUr4FssQ8.twitter

human use dataset

Seems like IUCN data is a good plan for this. but might need some parsing

google trends

google trends is transformed using log10(sum of hits +1)

human use

group

group the mammals into smaller groups
check how they split
e.g. terrestrial vs marine, eutherian vs non-eutherians, etc.

what to do with missing data in weight

@itchyshin says potentally we impute using phylogeny etc - https://github.com/ericgoolsby/Rphylopars/wiki

where to add h-index / altmetric / google trends?

should I have the h-index / altmetric / google trends data in another dataframe? or should I add them to the one with mass / phylogeny / etc.?

check gbif data

check what gbif is doing with domesticated animals

IUCN analysis

ordinary variable for IUCN status LC, VU, EN, CE, EW
other categories set to NA

check synonyms

check if there are any more synonyms

list of datasets and sources

• phylogeny

• body size data

• GBIF

• taxonomic synonym matching

pkgdown

@jessicatytam - is this something you may consider

https://www.tidyverse.org/blog/2020/09/pkgdown-1-6-0/

I realise that Will is not on the list - invite him over again

workflow diagram for proposal

possibly following #4

combining list for the 3rd time

Ok so i thought i did not match the synonyms correctly, which removed some important spp from the list (e.g. sheep).
Turns out the problem is with the rotl phylogeny list, which didn't have some of these spp and i removed spp that were not on this list.
So i am now making a new list keeping spp not on the rotl list, and getting h-index and google trends again.

conservation keyword

species name + "conservation" for all species
domesticated species higher h-index difference
use a bar plot
find out if they are partially / fully domesticated (add a column)
default is "not domesticated"
https://en.wikipedia.org/wiki/List_of_domesticated_animals

scaling & models

need to use a model with log values
model with real or log values?

pivot_wider()

use pivot_wider() to tidy data

https://tidyr.tidyverse.org/reference/pivot_wider.html

phylogeny

remove extinct species

• wooly mammoth
• homo genus

check on rotl

human use

change categories to binary
find the slope

run specieshindex on species list

also after #4

phylogenetic signal

get branch lengths from nature paper

internet popularity

this seems promising: https://cran.r-project.org/web/packages/gtrendsR/gtrendsR.pdf

get species centroids from gbif data

after #4

body mass

domestication

synonym matching

try taxize and rgbif

domestication

change categories to 1:3 (add column)