I've run CAT6 throughout my house, mostly in the basement where the TV area and office are. I have a cable that goes upstairs to my bedroom along the chimney from the basement to feed a wireless access point. There are 3 spots where the cable comes about 4-6 inches away from electrical wires. The power and ethernet wires are parallel for about 18" and about 4" away from each other. Will this make an impact on my data signal?
CAT6, even "unshielded", is very resistant to electrical interference, just as it can carry very high-speed data while emitting little or no interference. Plus, your power cables have pairs of conductors carrying current in opposite directions, so any interference they emit is going to rapidly diminish with distance.
(For fun, here is an explanation of how CatX cables reduce crosstalk and increase EMI resistance.)
In real terms, 4" is plenty of spacing, even if you ran the two lines parallel for much longer distances.
NFPA 70, National Electric Code (NEC) Section 800.133(A)(2) requires communications wires and cables shall be separated by at least 50mm (2 INCHES) from conductors of any electrical light, power, Class 1 non-power-limited fire alarm or medium-power network-powered broadband communications circuits. Exceptions are if separate raceways or conduit are used for separating the communications cables/wires from the power conductors.
This is for parallel runs. They can cross perpendicular at less than 2". An electrical inspector can fail a permit inspection and keep you from obtaining a certificate of occupancy (CO).
Differential Manchester Encoding generally prevents induced voltages on your Ethernet cable from power lines from creating an issue. Because he voltage induced on both lines of a twisted pair is the same the differential is zero.
The NEC standard is 2" and that is valid for 120/240V; however, NEC is only concerned with voltage induction, protecting the gear at each end and not starting fires.
At high bandwidth, you need to consider electromagnetic and radio interference. There is nothing worse than trying to debug Packet loss down to a layer 1 (physical issue).
Two things that you need to consider: Termination and throughput. Termination: It doesn't matter if you have shielded, top quality plenum if you leave your (un)twisted pairs exposed out the back of your RJ45.
Throughput: At 1Gbit or less 2-4" (separation from parallel power cabling) is sufficient for data over shielded (FTP) cable.
For unshielded (UTP), you should try for 8-16" separation.
Having metal conduit separation of either the power or the ethernet will allow you to use the lower bounds.
Crossing at right angles is fine.
At higher bandwidth, double the distances and make sure you ground your FTP.
Remember that interference is situational depending on the current/load you are putting on your power line and the amount of data that you are trying to transfer. This is why a layer 1 packet loss issue is so horrible to nail down.
A bad job is going to be noticeable at even 100Mbps. So when you call your fiber provider and complain, they are going to tell you that they clock UmptyBazillion-bps to the customer prem equipment and you are both going to leave the call frustrated.