Missing variables

[sBrook3]

Hello,

Thankyou for building and sharing this, its been very useful. I think probably the issue I'm having is lacking experience with R, but I've tried reading a number of different fit files (from different types of bike computer), and in all of them I'm seeing a much reduced range of variables than I can see when converting them to csv first or viewing through eg. training peaks. I can do without most of it but unfortunately torque is missing and I need that for what I'm trying to do.

Here's what I'm doing:

new_ride<-readFitFile('/Users/sarah/Downloads/new_project.fit') 
new_ride_records<-records(new_ride)%>% 
                  bind_rows()%>% 
                  arrange(timestamp)

which gives me

str(new_ride_records)

## tibble [7,464 × 20] (S3: tbl_df/tbl/data.frame)
##  $ timestamp                   : POSIXct[1:7464], format: "2024-03-21 15:31:20" "2024-03-21 15:31:21" ...
##  $ position_lat                : num [1:7464] NA 51.7 51.7 51.7 51.7 ...
##  $ position_long               : num [1:7464] NA -0.0629 -0.0629 -0.0629 -0.0629 ...
##  $ gps_accuracy                : int [1:7464] NA 7 7 7 5 5 5 5 5 5 ...
##  $ altitude                    : num [1:7464] NA NA NA NA NA NA NA NA NA NA ...
##  $ distance                    : num [1:7464] NA 0 5.57 11.67 17.65 ...
##  $ heart_rate                  : int [1:7464] 99 100 100 100 99 99 98 98 98 98 ...
##  $ calories                    : int [1:7464] 0 0 0 0 0 0 0 0 0 0 ...
##  $ cadence                     : int [1:7464] 0 2 82 83 84 86 86 86 87 51 ...
##  $ speed                       : num [1:7464] NA 5.85 5.92 6.1 5.89 ...
##  $ power                       : int [1:7464] NA NA 158 133 110 104 90 82 69 34 ...
##  $ left_right_balance          : int [1:7464] NA NA 57 58 55 54 49 54 50 56 ...
##  $ combined_pedal_smoothness   : num [1:7464] NA NA 51.5 50 49.5 51.5 55 47 46 48 ...
##  $ temperature                 : int [1:7464] NA NA NA NA 19 19 19 19 19 19 ...
##  $ grade                       : num [1:7464] NA NA NA NA NA NA NA NA NA NA ...
##  $ ascent                      : int [1:7464] NA NA NA NA NA NA NA NA NA NA ...
##  $ descent                     : int [1:7464] NA NA NA NA NA NA NA NA NA NA ...
##  $ battery_soc                 : num [1:7464] NA 82 NA NA NA NA NA NA NA NA ...
##  $ total_hemoglobin_conc       : num [1:7464] NA NA NA NA NA NA NA NA NA NA ...
##  $ saturated_hemoglobin_percent: num [1:7464] NA NA NA NA NA NA NA NA NA NA ...

and the session Info:

sessionInfo()

R version 4.3.0 (2023-04-21)
Platform: aarch64-apple-darwin20 (64-bit)
Running under: macOS 14.4.1

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib 
LAPACK: /Library/Frameworks/R.framework/Versions/4.3-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.11.0

locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8

time zone: Europe/London
tzcode source: internal

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] dplyr_1.1.4    FITfileR_0.1.8

loaded via a namespace (and not attached):
 [1] bit_4.0.5         compiler_4.3.0    promises_1.2.0.1  tidyselect_1.2.1  Rcpp_1.0.10       leaflet_2.2.2     later_1.3.1       yaml_2.3.7        fastmap_1.1.1     mime_0.12         R6_2.5.1          generics_0.1.3   
[13] knitr_1.43        htmlwidgets_1.6.2 tibble_3.2.1      shiny_1.8.0       pillar_1.9.0      rlang_1.1.3       utf8_1.2.4        httpuv_1.6.11     xfun_0.39         bit64_4.0.5       cli_3.6.2         withr_3.0.0      
[25] magrittr_2.0.3    crosstalk_1.2.0   digest_0.6.31     rstudioapi_0.14   xtable_1.8-4      lifecycle_1.0.4   vctrs_0.6.5       evaluate_0.21     glue_1.7.0        fansi_1.0.6       rmarkdown_2.22    tools_4.3.0      
[37] pkgconfig_2.0.3   ellipsis_0.3.2    htmltools_0.5.5  

Thanks again, and any suggestions you can offer would be much appreciated! Sarah

[grimbough]

Can you share a file? In my experience FIT files contain quite a bit of undocumented data, and I choose to throw away anything that I don't know what it is. I think the most likely thing is that torque is somehow being discarded internally, which I should be able to fix with an example file to debug.

[sBrook3]

new_project.fit.zip It wouldn't let me upload a .fit directly so hopefully that will work.

thanks

[grimbough]

Thanks for that file. Can I ask what you use to convert to CSV? I've tried a few tools like Garmin's FitCSVTool and the perl fitdump tool, but there doesn't seem to be any mention of torque in the files those tools create.

[sBrook3]

I'm using this: https://bluecattechnical.shinyapps.io/EnDuRA/ you have to create a login but its free.

[grimbough]

Thanks for the link. My guess is that they are calculating the torque value using the power and rpm data that are already in the file using the following equation:

Torque (nM) = Power (W) ∗ 60 / (2 ∗ pi ∗ RPM)

Here's some code that get's a plot similar to that produced by EnDuRA:

library(FITfileR)
library(dplyr)
library(ggplot2)
library(lubridate)
library(hms)

## download and extract the example file
destfile = tempfile(fileext = ".zip")
url = "https://github.com/grimbough/FITfileR/files/15011268/new_project.fit.zip"
dl = download.file(url = url, destfile = destfile)
example_file = unzip(destfile, exdir = tempdir())

## read and extract the records
fitFile <- readFitFile(example_file[1])

laps <- laps(fitFile) %>%
  bind_rows() %>%
  arrange(timestamp)

lap_intervals <- interval(laps$start_time, laps$timestamp)

records <- records(fitFile) %>%
  bind_rows()%>% 
  arrange(timestamp)

records <- mutate(records, 
                  ## find the time elapsed since the 
                  elapsed_time = as_hms(timestamp - min(timestamp)),
                  ## calculate the torque at each second
                  torque = (power * 60) / (2 * pi * cadence),
                  ## assign each record to a lap - this is a bit slow
                  lap = as.factor(
                      sapply(timestamp, FUN = \(x) which(x %within% lap_intervals)[1])
                    )
                  )

## create a plot
ggplot(records, aes(x = elapsed_time, y = torque)) +
  geom_line(aes(color = lap)) +
  scale_x_time() +
  theme_minimal() +
  ylim(0,100)

[sBrook3]

Thanks so much for digging into that. I'd completely assumed the torque would be coming directly from the power meter. I've tried it on a few rides and it seems to work.

Thanks!