TechNomad & Elepa
TechNomad TechNomad
Elepa Elepa
TechNomad TechNomad
I’ve been mapping my daily commute times in each country to a simple line graph—any thoughts on cleaning up outliers?
Elepa Elepa
If the outliers are just a few errant spikes, just trim the top and bottom 5 % of the data and recompute the trend line; otherwise, try a median‑based moving average. That’ll keep the line stable without over‑fitting to noise.
TechNomad TechNomad
Sounds solid—just a quick sanity check on the trimmed data before I push the new graph to the team. Thanks for the quick fix!
Elepa Elepa
Run a quick descriptive stats check on the trimmed set—mean, median, standard deviation—and compare those to the original; if the variance drops below your threshold, you’re good to go.
TechNomad TechNomad
Got it—just calculate the mean, median and standard deviation of the trimmed set, then compare each value to the same metric from the full data. If the variance (or standard deviation) drops below your threshold, you’re set to proceed.
Elepa Elepa
Glad the plan works—just remember the trimmed stats should still be within a few percent of the originals, otherwise you might be hiding a real pattern. Happy graphing.
TechNomad TechNomad
That’s the move—nice reminder to keep an eye on the shift. Got my laptop set up on a hammock in Bali, so I’ll crunch those numbers over a cup of kopi luwak and hit you back with the graph. Happy coding!
Elepa Elepa
Enjoy the Bali breeze, but keep a log of your coffee breaks—caffeine spikes can produce temporary outliers, so a small sliding‑window filter might keep your mean stable. Happy crunching.
TechNomad TechNomad
Will do—coffee log going into my notes, and a quick sliding‑window filter to smooth out any caffeine‑driven spikes. Thanks for the tip, keeping the numbers honest while I enjoy the breeze!