Can we guess sport/activity for TCX from stats?

[mlt]

Unfortunately there is no way to tell ahead what that track is all about. It might be possible to deduce it for a particular user from one's track statistics. If median speed & IQR are somewhat high, we are probably biking, otherwise running.

Here is the pic with some stats. Red line results from natural breaks (jenks) split of mean speed into 2 classes.

library(ggplot2)
library(classInt)

root <- "~/Documents/My Runs"
trk.lst <- list.files(root, "*.track$", full.names=TRUE, recursive=TRUE)

trk <- lapply(lapply(trk.lst, read.csv),
              function(x) {
                  x["x1"] <- NULL
                  x
              })
trk.all <- do.call(rbind, trk)

s <- with(all, Speed[Speed>0])
c <- classIntervals(s, 2, "jenks")
threshold <- c$brks[2]

pts.lst <- list.files(root, "*.points$", full.names=TRUE, recursive=TRUE)
pts <- lapply(pts.lst, read.csv, colClasses=c(Track="factor"))
pts.all <- do.call(rbind, pts)

ggplot(pts.all, aes(x=Track, y=Speed)) +
    geom_boxplot() + stat_summary(fun.y=mean, geom="point", shape=23) +
    geom_hline(aes(yintercept=threshold), colour="red") +
    theme_bw()