blink summarizes your data tailored to the need of
ophthalmic research: It looks for VA and IOP columns and summarises those
with common statistics. In order to make it work, it requires specific
column naming - please see section "column names" and "data coding".
For more details how blink works, see vignette("eye")
blink(x, va_to = "logmar", va_cols = NULL, iop_cols = NULL, fct_level = 0:4)object of class blink and list. Class blink contains the
myopized data, count of patients and eyes,
and summaries for visual acuities and intraocular pressure.
data frame
to which VA notation (passed to va())
if specified, overruling automatic VA columns selection. tidyselection supported
if specified, overruling automatic IOP columns selection. tidyselection supported
Remove columns for Summarizing when all unique values
fall into range. character or numeric vector, default 1:4
Only common codes supported:
eyes: "r", "re", "od", "right" - or numeric coding r:l = 0:1 or 1:2
Visual acuity: "VA", "BCVA", "Acuity"
Intraocular pressure: "IOP", "GAT", "NCT", "pressure"
No spaces!
Do not use numeric coding for eyes in column names
Separate eye and VA and IOP codes with underscores ("bcva_l_preop", "VA_r", "left_va", "IOP_re")
Avoid separate VA or IOP codes if this is not actually containing VA/ IOP data (e.g. "stableVA" instead of "stable_va", ChangeIOP instead of "change_IOP")
Keep names short
Don't use underscores when you don't have to. Consider each section divided by an underscore as a relevant characteristic of your variable. ("preop" instead of "pre_op", "VA" instead of "VA_ETDRS_Letters")
Use common codes for your patient column (see eyes, section Guessing) (e.g., "pat", "patient" or "ID", ideally both: "patientID" or "patID")
Don't be too creative with your names!
Good names:
-c("patid", "surgery_right", "iop_r_preop", "va_r_preop", "iop_r", "iop_l")
OK names
-c("Id", "Eye", "BaselineAge", "VA_ETDRS_Letters", "InjectionNumber"):
Names are long and there are two unnecessary underscore in the VA column.
Better just "VA"
-c("id", "r", "l"): All names are commonly used (good!),
but which dimension of "r"/"l" are we exactly looking at?
Bad names (eye will fail)
c("id", "iopr", "iopl", "VAr", "VAl"):
eye won't be able to recognize IOP and VA columns
c("id", "iop_r", "iop_l", "stable_iop_r", "stable_iop_l"):
eye may wrongly identify the (probably logical) columns "stable_iop"
as columns containing IOP data. Better maybe: "stableIOP_l"
c("person", "goldmann", "vision"): eye will not recognize that at all
blink and myop work more reliably with clean data (any package will, really!). clean data.
Done with remCols: Removes columns that only contain values defined in fct_levels or logicals from selected columns (currently for both automatically and manually selected columns). fct_levels are removed because they are likely categorical codes.
blink is basically a wrapper around myop, eyes and reveal:
Duplicate rows are always removed
Column names are prepared for myopization (see myop)
VA will always be converted to logmar
How to rename your columns (two threads on stackoverflow.com):
library(eyedata)
blink(amd2)
messy_df <- data.frame( id = letters[1:3],
iop_r_preop = sample(21:23), iop_r_postop = sample(11:13),
iop_l_postop = sample(11:13), iop_l_preop = sample(31:33),
va_r_preop = sample(41:43), va_l_preop = sample(41:43),
va_r_postop = sample(51:53), va_l_postop = sample(45:47)
)
blink(messy_df)
Run the code above in your browser using DataLab