Help interpreting oxygen sensor data

Hi,

After seeing a decrease in mileage on my 96 Ford Taurus wagon (~ 24 to
~20 mpg. 140.000 Mi odo), I decided to take a look at my oxygen sensors
as a possible culprit.
I read out their voltages during idle when warm using my Elm ODB
scantool, and plotted them.

http://www.duijn.info/OxygenSensor.jpg
x in seconds, y in volts

It looks to me that something is fishy in Bank 2, but the signal does
not look like the typical error modes that I read about (sluggish, low,
flat). Bank 1 looks like the examples I saw for good sensors.

I am not experienced enough to make a statement on these curves, but am
hoping they ring a bell with one of you people. Does it look familar, is
my data aquisistion simply too slow (I don't have a scope, sorry...).
The peaks in sensor 2/2 also look like they don't belong there. Is the
fuel injection acting up due to bad info from sensor 1, and too rich a
mixture making its way through the cat? The traces were recorded
separately to improve time resolution...

If a sensor is indeed bad, is it doable to replace it with just basic
tools, or is it best left to a pro?

Thanks for any thoughts you may have.

Martijn