Chapter 2: Finding and wrangling time series data

Basically, anything that you can map to time is time series data. Timestamps can make this easier, but we have to be aware of what the timestamp represents, like when an even occurred or when the event was logged.

Dealing with missing data has to be done with consideration for the domain. Forward filling (just sticking the last value in there) is a quick thing to try, but rolling means or medians, or model driven approaches may be better. Especially consider model driven approaches if you can add process insight to the model. This lets you inject a prior into the system!