Valid weekly weigh-in is always useful.
As a woman, you need 4 weeks to get any useful data you could find useful though, as your BMR literally changes through the month.

For any weigh-in following the start of a deficit, make note, because should be bigger water weight drop.
Same as weigh-in coming out of a diet, bigger water weight gain.

Another good reason for a reset to be 2 months minimum, allows seeing if TDEE estimate for that level of activity is close.

Then again, may not have 2 months that are anywhere near being close to the same activity level - so you can only estimate and hope you got it on the high side, noting a mere 100 extra daily over true maintenance would take 35 days to show a slow 1 lb gain.

Valid weigh-in to minimize known expected water weight fluctuations -
Morning after rest day eating normal sodium levels, and not sore from last workout.

Sometimes that means you don't even waste time with a weigh-in (why did I have Chinese for lunch yesterday on rest day!), or you change the weigh-in day (gotta do my lifting workout for makeup).

You are manually logging your strength workouts right, since HR based is inflated?
I can't remember who catches that fact or not.

But keep an eye on your corrected Fitbit TDEE, since it could change and obviously a 5 level table isn't except by big gaps.

Daily is fine as long as it doesn't actually cause you to react in any way, emotionally (stress) or physically (eating less).
I'd just suggest tagging the valid weigh-ins so you can do math with numbers later if need be.

Otherwise, you almost really need 2 months of data to see a trend in all the noise.

Fitbit does have better database overall, or actually, they both start with the same database. But MFP converts to weight based formula, Fitbit leaves it at METS, or your BMR x activity factor - which is more accurate.
Fitbit also has a tad more entries that have intensity options - like elliptical or swimming.

I'd say even 4 week rolling avg is fine. One odd lazy week or totally active week won't throw it off too much, but you also don't have too long a lag time as seasons or daily life changes. And even there, you may see things changed big time and just go for 2 week average.

I know going in to winter is that way. One week willing to ride the bike down in the 40's and still some long runs, and bam, snow comes and bike is done, and depending on how much running is done or shorter. Massive change to TDEE for me when that happens.

If this was based on TDEE table, you'd eat back anything extra you didn't account for in your estimate of rough TDEE level.

Since using Fitbit average though, if you did a whole lot more than the usual you know the Fitbit is based on, then it too should be eaten back.

So in an average week, you might look at the high and low days. It could be your average 2400 is based on 1 high day at 2800 and 2 low days at 2200, other days hover around 2400. So with that being the basis of the 2400 - no need to compensate for that 2876. Unless that was the 2nd day of hitting it.

But if average week has highs of half 2500 and lows half 2300, then that 2876 truly was extra compared to planned and normal.
I'd suggest 2876 - 2500 = 376 overage less 15% = 320 to eat back on that day.

So really, just from stats from the Fitbit email, have in mind your average big days.
If you know you truly did more than planned (based on past average activity), then eat back with same deficit.

Because if you think about it - even those using the TDEE table are doing the same thing.
Literal TDEE on exercise days is bigger, but their eating level is still based on average of the week, so the deficit in reality is bigger than 15%. Same on rest days, depending on workout schedule, could have little to no actual deficit on rest days.