Less support question and more just curiosity, but for work I pretty frequently am at a single location for an entire day, but end up walking a few miles between traveling between different areas of the same building or location. I’ll attach pictures of what I mean, but just wondering if most people merge, delete, or leave it as is.
I’m guessing there’s no real answer.
I like merging them by marking these short walks as stationary. The only exception is if Google Places have separate names for different parts of the building.
Personally I most often end up adding new private places for each building or room within. Like if it’s an airport then I add places (or use existing Google Places if available) for each gate, for the shops, immigration, baggage claim, etc.
So my airport timeline might be: Taxi Drop-off Area → Check-in Area → Immigration → Lounge → Gate 123, with walking items between.
I sometimes also do like @trackerminerfs does - converting some walking to stationary, so that it’ll merge together into a single visit. But because that loses detail I’d often rather keep, I do it less commonly. But sometimes I really do want it as a single visit, so there’s no way around it other than converting those walking bits to stationary.
We do have a couple of issues filed for potentially improving this stuff:
Hi @twilldre,
I’ve tried that before, too: adding extra private locations and tagging my visits to them. But it’s so annoying to have to confirm (or rather, correct) Arc every single time. Because in my experience, Arc doesn’t learn that, for example, after I arrive and climb several flights of stairs (!), I’m not on the parking deck anymore / for six hours straight, but instead spend the rest of the day at the nearby office— no matter how often I correct it.
I know it’s hard to get good reception inside buildings, but I’m annoyed by all these glitches or “ghost signals” that happen almost every day, even though my cell phone is sitting on my desk or I’m just moving from one room to another on the same floor.
At an airport or on a campus, where the various areas are likely farther apart, this might work better.
Here’s another example—or an extreme measure: Two weeks ago, I turned Arc off completely for several days. I was on a very small island and, for nature conservation reasons, was confined to the house and its immediate surroundings most of the time. At first, I tried the approach described above (last year, I had added various private spots in front of and behind the house, as well as nearby areas in the garden where I spent anywhere from a few minutes to several hours throughout the day). I had hoped that Arc4 would track these different areas—or even my daily walks to two different spots on the coast that are a ways from the house—better than last year’s version did, but that wasn’t the case. Or it would activate so late that I’d already been out and about on the island for 15 minutes, but a large portion of that activity was already missing. I know that Arc needs time to “get to know” completely new areas… but aren’t even two weeks enough?
Oh, this has turned into a bit of a rant. I’m sorry about that.
My conclusion: I try to keep my visits as precise as possible by adding additional private locations when it’s important to me to know that I was at different spots within the same place or that i walked for a greater amount of time between visits. Otherwise, I don’t go to that much trouble and try to combine everything into a single, fixed-location visit.
Hi @midor — no need to apologise for the rant; this is exactly the kind of real-use detail that’s useful to us. The things you raised sit quite differently from each other, so let me sort them out:
Different places at the same spot (the parking deck below, the office above): Arc’s places are latitude/longitude only — there’s no altitude, and even if there were it wouldn’t be fine enough to tell floors apart. For same-coordinate places it leans instead on the pattern of your visits — mainly time of day, and how long you stay. Where two places have clearly different rhythms — a café you drop into in the mornings versus a gym you spend afternoons at — that works well. Yours is a harder case, though: you reach the parking deck and the office within minutes of each other each day, so there’s little timing difference to tell them apart on. Visit duration can still re-correct it later, once Arc re-evaluates unconfirmed assignments — but right at arrival, when both visits are only a few minutes old, there isn’t yet much to go on.
“Two weeks isn’t enough” for new areas: it’s less about elapsed time than about repetition with a consistent pattern. A handful of visits can be plenty if your arrivals and departures are regular — but if the visits are irregular (as travel and one-off spots tend to be), there’s no pattern to match, and it can’t really learn them however long it has. New, varied places are genuinely the hardest case, and a holiday is close to the worst.
The indoor “ghost signals”: those are location drift — your reported position wandering while you’re actually still. There’s a system specifically for suppressing it, and “it’s not catching mine well enough” is fair — but it’s something we can only act on with data. If you’d like us to look, the best path is a new thread with a couple of screenshots and a short description of a typical case, and we’d dig in from there.
The ~15 minutes missing on the island: this one’s different — that’s not expected. A minute or so as recording wakes up can happen, but 15 minutes is a genuine recording gap, not normal latency. If it recurs, a session log covering it (emailed to matt@bigpaua.com, same as the logs you’ve sent before, with roughly when it happened) would give us something concrete to chase.
No pressure either way — if it was mostly venting, that’s completely fine. But if any of these are worth pursuing for you, the drift one (screenshots, new thread) and the missing-recording one (a log) are where we could actually take it further.
