Ask me about my week on Friday - my answer depends entirely on what I'm doing that night. Out with the guys - it was a great week! Going home to do laundry - the week sucked. Way too many confounding variables in an emotional survey like this?
Its been shown in study after study, that people's response to surveys is extremely prone to suggestibility and situation. I remember a racial stereotyping survey with a movie afterward, where people who were told the movie was cancelled recorded as far less tolerant.
I like OP's idea a lot, however it will suffer from memory and social conditioning bias. This is why in Chronicle[1] I record these data points at the very moment they happen, and use the notion of "fold" to summarize them at daily/weekly/monthly level.
Its been shown in study after study, that people's response to surveys is extremely prone to suggestibility and situation. I remember a racial stereotyping survey with a movie afterward, where people who were told the movie was cancelled recorded as far less tolerant.