Data gaps and altitude spikes

Hello OpenSky, I am very new here so forgive me if I missed a FAQ. I perused the older posts and didnt see anything relevant to my question(s).

I found my way to the Air Traffic data samples and I'm in full delight mode for having such a great repo of interesting data. I am working with a full days worth of data downloaded from here: 
My use of this data is limited to just wanting to make pretty data-vis outputs. I have been working on a full featured full globe setup in Houdini, and I often look for data sources to map to the globe to produce interesting renders. Usually I do things like cities, roads, rivers, and add data like earthquakes or cell towers or airports. Here are a couple projects to give you an idea of the types of things I work on: EDIT: Oops, links arent allowed. Google Robert Hodgin taxi taxi to see the yellow-taxi data vis I did.

First question: I see occasional spikes in the altitude being reported. As a general curiosity, is this commonplace for dealing with airplane data? I dont fully understand how the data is being collected, so it might just be a side-effect of the process of collecting the data from a collective? 

Second question: Are there resources or processes that have already been discussed here for fixing these anomalies? I tried walking each flight path to look for altitude changes that exceed a threshold, and then interpolating between the correct values on either side of the glitch, and that works well for most of the issues I'm running in to. However, I am noticing now that many of the flight paths lose time or have the plane at the exact same lat/long for minutes before it jumps to its correct location. These are proving to be harder to mend. Im just curious if you have come across anyone who has managed to patch the data in a reasonable way.

Third question: The readme states "You'll find one state vector per second for each aircraft which was active within the coverage of OpenSky at that particular second." In the data I downloaded, it looks like there is a data point every 10 seconds instead of every second. Not an issue for me since every 10 seconds is plenty of data to keep me busy. I am just curious if this is a typo? Or maybe my assumptions about the data time stamps are incorrect.

Finally: Knowing what Im up to, are there any other data sources you might recommend me take a look at?

Regardless, this resource is fantastic and many thanks to folks that are keeping it up and running. I look forward to being able to share my outputs with you.
-robert