Previously, the 'latin1' encoding causes most data of Ishihara0000 to be not read in at all. Check encoding settings elsewhere.

The author replied the same thing for both studies:
6. Is the 'Stand description' complete?
If not, could you please provide more information? Codes and legends for Growing_condition and Status are found within Variable.definition.csv file attached.
ANSWER: status a primay hill dipterocarp forest

I have added a new_question.txt file asking for it be fixed according to our standards.

On the reply file 'report1-questions.txt' the author recognised that some data had incorrect units (I corrected those already) others are outliers. What do you want to do?
See reply below:
9. Please review the plots showing how your data compares to other data in the study. Does your data fall well within the rest of the dataset? Are there outliers?
If so, could you please review the original data file you provided us with and verify that all information is correct.
ANSWER: The crown width of some individuals and 2 LMA value were out of fit. These data were not in the right unit and it was corrected in "Cor_data_delagrange0000a.csv". Others are outliers.

In previous reports, it says we have 14152 measurements, but we now have only 13671, although I don't recall deleting anything?

I looked at the number of observations in the raw data file, vs what shows up in the final dataset (code further below).

Results:
study nmiss
1 Coll2008 132
2 McCulloh2010 348
3 O'Grady2006 10
4 Osada0000 184
5 Osada2003 223
6 Osada2005 49
7 Peri0000 12
8 Peri2008 9
9 Peri2011 9
10 Selaya2007 464
11 Selaya2008 535
12 Selaya2008b 263
13 Wang2011 4

NOTE: McCulloh2010 needs to have many missing ones, because all samples were branch-level, only some were tree level. We have to go through the other ones and check!

dat <- loadStudies()$data
studynames <- unique(dat$dataset)
rawdat <- lapply(studynames, readDataRaw)

nstudy <- as.vector(table(dat$dataset))
nstudyraw <- sapply(rawdat, nrow)
nmiss <- nstudyraw - nstudy
df <- data.frame(study=studynames[nstudyraw > nstudy],
nmiss = nmiss[nmiss >0]
)
df

The author replied

1. I see in your paper that you provide an average value for canopy openness [average 5.9 +- 0.7%], however we did not get individual information. Would you be willing to share individual-based info in case you have it?
   ANSWER: Canopy openness is available only for three individuals per species and therefore would be useless for purpose of a meta-analysis.

I followed the author's advice and removed this information. I did this in a separate commit on purpose in case you want to go back and keep it in the final dataset.

Are they in the sea?

Test code:

source('report/report-fun.R')
dat <- loadStudies(reprocess=FALSE)
 prepdata  <-  prepMapInfo(dat$data)

Dan, could you please edit the bibtex reference for these studies? One (2008) is a proceedings abstract in spanish and the other (2011) seems to be a book chapter.
The full references provided by the author can be found within the review folder.

Guys, we got the following reply from the author

1. Do your locations fall in the right spot in both world and country map?
   If not, please outline the issues here and provide us with updated longitude and latitude data.
   ANSWER: MAP IS CORRECT BUT LONG SHOWS NEGATIVE AND IT SHOLD BE POSITIVE

The longitude we have is -123.6556, which corresponds to California in the US. So I don't see any problems? Why does the author think that it should be positive?

Sending email to the following people, based on code as of commit e685da2

rm(list=ls())

source('report/report-fun.R')
dat <- loadStudies(reprocess=TRUE)

emailReport(dat, "Lusk2012")
emailReport(dat, "Kelly0000a")
emailReport(dat, "Baltzer2007")
emailReport(dat, "Bond-Lamberty2002")

The author provided an 'in review' reference:
O'Hara, K.L., York, R.A. (in review). Leaf area and crown architecture in a giant sequoia spacing study: A new model for LAI development. Forest Science

what do we do?

During review Remko and Daniel identified number of variable name changes, in column in variable definitions file. Need a function to find and replace a variable name throughout project. Untested draft

    fn <- list.files(pattern="\\.R", recursive=TRUE, full.names=TRUE)
    for(i in 1:length(fn)){
       r <- readLines(fn[i])
       r <- gsub("raw","rawdata",r)
     writeLines(r, fn[i])
    }

For studies in which a certain variable (e.g., 'traits') was not measured, how do we treat it?
Keep as an empty cell, call it NA, N/A, or delete that row?

• Conifer leaf area - define ideal, then implement it.
• Definition should be 'half of total surface area' (other measurements often include either 'projected area', or 'total surface area')
• check all studies with conifers to standardise measurements

1. Please review the plots showing how your data compares to other data in the study. Does your data fall well within the rest of the dataset? Are there outliers?
   If so, could you please review the original data file you provided us with and verify that all information is correct.
   ANSWER: for some surface area plots, the data are clear outliers, but apparantly because similar combinations of traits were not available for studies with large trees (and large surface areas....)?

The author did not provide a metadata, I have added a new_question.txt file requesting to do so.

dat <- loadStudies(reprocess=TRUE)
str(dat$data)

note the
..- attr(*, "names")= chr NA NA NA NA ...

for a.ilf and m.ilf
This is a side effect of a data processing problem?

The author provided incomplete information
6. Is the 'Stand description' complete?
If not, could you please provide more information? Codes and legends for Growing_condition and Status are found within Variable.definition.csv file attached.
ANSWER: STAND DESCRIPTION: Multiaged ponderosa pine stands on range of soils and plant associations.

I have added a new_question.txt file so he can re-check the answer based on variable definitions guidelines.

What do we do with this? dataMatchColumns indicates m2 anyway, but does this confer some weird comparisons?

1. It seems that the units you provided for leaf area might be wrong. You provided m2 but we suspect it may be cm2. Could you please double-check this information? The original indicates m2 indeed, but maybe the original dataset is expressed in cm2?
   ANSWER: should be m2, but should be specified as "all-sided leaf area"

Guys please check this out.. it will be easy to modify but we need to first arbitrarily define one status category for each stand description.

1. Is the 'Stand description' complete?
   If not, could you please provide more information? Codes and legends for Growing_condition and Status are found within Variable.definition.csv file attached.
   ANSWER: I don't quite understand what is meant by the codes dominant, codominant and suppressed in terms of a stand as they relate to trees, but in my understanding the sparse stand was all dominants, the medium and the dense had dominants and suppressed trees (now the medium seems to have the most dominant trees)

• Check methods variables for h.c
• Delete c.d variable, for these studies ensure h.c is calculated correctly (Aiba2005 Delagrange2004 Osada0000 Osada2003 Osada2005 Osunkoya2007 Petritan2009 Sterck0000)

• Check input files (make function to do this routinely, perhaps as
  the project files are updated).
  • order of column names should correspond
  • no dupicate column names
  • no NA values in var_in (e.g., was in O'Hara0000)
    This is harder than it looks because things like "Genus" are
    sometimes missing from the matching columns. Some work will be
    needed here.

using taxize or other method?

How do we define this description within our status categories?

"Saplings in understory and gaps of tropical humid forest on Barro Colorado Island, Panama"

The author said that the vegetation type was 'mesic temperate deciduous forest'
How does that fit into our existing categories? TempF or TempRF?

Check all studies with branch mass. Is definition consistent across studies?

The author indicated that one of the variables was mistakenly interpreted

1. Please review the plots showing how your data compares to other data in the study. Does your data fall well within the rest of the dataset? Are there outliers?
   If so, could you please review the original data file you provided us with and verify that all information is correct.
   ANSWER: WE DID NOT Have sapwood data at crown base so the one graph that shows our data as outliers is incorrect. We should have no data on that graph.

I'll leave to you to decide what to do with the variable a.ss2 (wrongly converted into a.ssbc)

Review instructions in readme file

Will close this when the author returns the correct file.

Want to manage which data folders are complete or incomplete within data folder.

Abstract has special characters and author's initials do not match the actual citation. How do we solve this?

Here's a possible solution that I tried for problem with authors' names:
http://www.tex.ac.uk/cgi-bin/texfaq2html?label=bibtranscinit

What do you think?

We originally added a question:

1. Is crown height equivalent to height to crown base, being the height of first branch bearing a live leaf?
   ANSWER: Yes.

I went back to check the data.csv file but can't see this variable there. Did we delete it at some point? If so, why?

1. There are 9 data points where 'Leaf mass (kg)' is zero. Can you please check whether these data are correct?

ANSWER: The data is correct. The data are from deciduous species that at the moment were without leaves.

Search on git custom diff files
possible solution?
http://git-scm.com/book/en/Customizing-Git-Git-Attributes#Binary-Files

I asked to author for this specific item already, just waiting for the reply.

Kantola provided location information for his dataset indicating lat, lon, status and vegetation type. These three locations however, are not linked to any particular individual, which makes it not possible to incorporate. I have added a new_question.txt file into the review folder so we can email him asking to add this information to the data.csv file. This means that the location-level information that he provided will ONLY be incorporated after we can assign each individual to a particular location.

I added a new_question.txt file asking for it.

I have created a new public repo called dataMashR (https://github.com/dfalster/dataMashR). The core data import functions (in R/import.R" should be moved over there and then used within this project as a package installed from github.

The author has provided the following answer:

1. Please review the plots showing how your data compares to other data in the study. Does your data fall well within the rest of the dataset? Are there outliers?
   If so, could you please review the original data file you provided us with and verify that all information is correct.
   ANSWER: It seems that 2 populations or units are used for leaf N concentration in the entire dataset.

We should write a function that compares two csv files and identifies the differences, so that we can see what has changed. Git does this, but only on a line-by-line basis. We want cell by cell.

Any ideas how to do this?

the library testthat can be used to compare to sv files, but only returns true or false,
library(testthat)
df1 <- read.csv(...)
df2 <- read.csv(...)
expect_that(df1, equals(df2))

Delagrange0000a has four values where h.t = 0