Hi @matt, have a great day. I am having an extremely problematic incident with Arc Timeline. Recently I have some issues with my iPhone that requires me to change the system date/time to a point in the past (3 August) to investigate. However, when I changed the date, Arc Timeline seems to continue recording current location data to this past date. And now, when I reset the system date to the current date, Arc has been creating a big mess, making every single day of the past four months (from 3 August till now) a travel segment that I don’t know how to fix. It appears on every single day of the timeline from 3 August till now a segment that says I am travelling. Is there any way that I can get this fixed? Thanks a lot.
Hi Matt. And now for some reason, my entire week of data has gone. Last week’s timeline shows nothing. And the data back in August, only the specific points of visit were there, all the travelling data (walking, train, metro, etc) has also gone too. What am I supposed to do? It is so frustrating…
Hi @matt, right now I have created a copy of the backup folder in iCloud Drive, in case the new data got backed up and overwrote the old backup files. However, I don’t know how to import these old files into the 3 August timeline entry. My idea is that I would like to overwrite the entire 3 August data with the older, backed-up version in iCloud Drive. Can you tell me how I can achieve this? Thanks a lot.
Oh dear! That sounds a really difficult one.
If you have backup data to edit and restore, then that’s probably the most conclusive fix, with the best potential results, but it will be incredibly difficult and fiddly editing all the files to correct the dates.
The other option is to find the bad data and mark it as “bogus”. Which will also be extremely fiddly, but likely less fiddly than all the JSON editing. Once samples have been marked as “bogus” they no longer contribute to timeline item calculations or processing - they’re essentially deleted (albeit with the option to restore them by changing them to some type other than bogus later).
If you go down the JSON editing route, you’ll want to correct the dates on each recorded LocomotionSample, and each TimelineItem, and also update the lastSaved dates to be recent, so that the importer recognises it as updated/newer data.
Importing the edited JSON can be done from the File Importer view in Settings → Backup, Import & Export. Copy the edited JSON files into the Import folder on iCloud Drive, then in named folders inside there (sample week files in a LocomotionSample folder, and timeline item files in a TimelineItem folder). The importer will then be able to see them, and will allow importing to update existing data.
But yeah, that’s going to be quite a fiddly process! Feel free to ask any questions you need.
Oh wait, I just remembered I think @more has a tool specifically built for doing those kinds of edits. Though my memory is terrible on the details.
Thank you for your prompt reply @matt. I have been extremely worried before hearing anything from you 
I think the JSON editing does not seem viable without the help of a tool. The weird thing is that I can’t seem to find the bad data via the individual segment view. Essentially, I don’t know how to find the exact ones. The software does not stop recalculating after I make an edit, which is very laggy and causes the app to terminate itself.
Do you mind if I email you screenshots of the problem? Cause there are some private data that I don’t want to post publicly,. And if you can find the other tool to help me with this case, that would be really appreciated. I don’t want to give up months and years of data on Arc, to be honest.
Now my timeline has a very weird segment that seems to encompass all my travel data for… 3 months and it appears in every single day on the timeline till then. Do you have any idea what this is all about? Thank you so much!
Meanwhile, the size of Arc’s backup in my iCloud Drive has doubled since the incident happened. I suspect that Arc has added a layer of bad data on top of the real actual one, but have no idea how to remove that.
I found the tool: GitHub - thoughtgap/arc-data-server: Data Wrangling with Web API for exported personal location records from Arc App https://bigpaua.com/arcapp
I haven’t used it myself, but have seen it in use a bunch of times and it looks very powerful. Though it might take some learning to get familiar with how to use it in this specific case.
Don’t worry, I’m sure your data is all there! Arc never deletes any recorded data, so it will just be a matter of figuring out how to edit it to undo the incorrect dates.
Or if we’re going the route of marking the incorrectly dated data as bogus, then I guess our goal in that case will be to “hide” that data (while it’ll technically still be in the database). I suspect it’ll be easiest to go that route first. Then once we’ve got as far as possible taking that route, we could work on finding that bogus data in the JSON files and editing the dates on it to hopefully restore it back to a sensible and usable state.
I’m not yet forming a good intuition for it, but I suspect it’ll be the processing engine having merged a bunch of things together due to overlapping in time, which has resulted in a “mega visit”, ie a single visit in the timeline that’s swallowed up a bunch of things that shouldn’t be inside it.
The challenge will be to edit as much of the problematic data out of that mega visit! Which might be quite challenging, if the visit has got so much data in it that Arc can’t load it into memory when editing from the Individual Segments view etc.
Oh wait, in your screenshot I see it’s a mega trip item! Ok, well, I guess that’s the same thing. The process won’t be any different when attempting to clean it up. Though a mega trip is slightly less problematic than a mega visit, because visits are prone to swallowing adjacent items into their gravity well, while trip items don’t do that. So the problem at least shouldn’t be getting worse over time. That’s some small comfort!
Anyway, basically you need to get to the Individual Segments view of that mega trip item (you might be able to get to it directly by long pressing on it on the timeline view, to get to the context menu, skipping the intermediate details and edit views).
Then once you’re on the individual segments view… The challenge is to find a suitable segment to edit, then try to split that segment, isolating segments that look like they have the bad dates, and marking them as bogus. Easier said than done!
Presumably the newer data in that mega trip will be correctly dated while the older data won’t be. Though I guess depending on how the processing engine tried to resolve it when the mess first started will define whether the incorrect dates are at the very beginning of the item or somewhere in the middle, or… Yeah, it’s going to take some experimentation.
Maybe it’s worth having a look at the Individual Segments view for that item (if you can get to it!) You can always email me screenshots at matt@bigpaua.com if you don’t want to post them here.