Handle Unicode escape sequences for emojis and symbols

[ShortTimeNoSee]

Currently when a data export has emojis that are in the form of Unicode escape sequences (e.g., \u00f0\u009f\u0098\u00ad), it is not handled correctly on analysis and causes incorrect statistics for emoji data.

[ShortTimeNoSee]

There is also an issue with symbols

This returns as "donâ" when it's supposed to be "don't" (except their apostrophe is the curvy type, so "don’t"). So we get stuff like this as a result lol: