Dates are categorical?
Why are dates categorical? They have numerical value. You can add or subtract them, and calculate their mean.
As per me: Because, we cannot have meaningful stuff in business perspective out of doing manipulation on dates. Like what would be use of calculation of mean of dates as per you? Same was thing with customer id in real-life example of traditional data (previous video). As it was sequenced, ordered and numerical number but doing mathematical manipulation on it won't led to any useful result.
Hi. "What would be use of calculation of mean of dates as per you? Suppose I have aggregate insurance data by policy year, and I want to apply an inflation factor from the average date of the policy year to today. Suppose that policy year is from October 1, 2021 to September 20, 2022. I would think taking an average of those two dates IS meaningful.
Sorry, I naively gave my pov🥶 . Sorry for incorrect explanation if it is. Let's wait and hope soon instructor provides correct answer to the same.
Hi Bhaskar and fcas80!
Thanks for reaching out.
In fact, it depends what operation we are referring to.
If the idea is to only store data and relate its values to specific dates, then 'date' in itself is a categorical value. It can have no numeric meaning, since it indicates a moment in time.
Instead, when, as fcas80 suggests, we want to measure the distance in time in terms of number of days, weeks, months or years, for example, 'date' can be treated as a numeric value. We usually store a date value as the 1st of January 1970 plus the number of days that have passed since then (unix time).
Therefore, we can convert from one to the other type in order to serve our purposes, but in our lecture, we present a date conceptually as a categorical value.
Hope this helps.