Email column

[sebastian]

There is a column called email_please where people are free to provide their own email address.
Currently it's null for all users.

Additionally we have an id column which has a unique id provided by typeform.

I suggest:

• rename email_please to email
• we use the email column as the uid for purposes of anonymization
• in the ETL (data transformation step when the data is mungded we) we use the id value for the email column when the email address hasn't been provided
• if an email has been provided then that is used

This setup allows us to capture cases where someone provides multiple survey answers and provides the same email address, and secondly also ensures we anonymize with respect to the email address, and as such never leak it.

[yoid2000]

As a temporary step, I check for email address, and if there is one I add some randomness to the uid.

This is just a safety check.

[sebastian]

But that doesn't help at all! That way the UID is guaranteed to be unique, which means the emails don't end up protected at all.

[yoid2000]

Ah sorry. I also make the email column NULL. Forgot to mention it.

(Though now I'm not sure why I bother to make the uid unique, since it already is...)

Anyway I'll take care of the rest of this issue later. Low priority right now.