Database Reference
In-Depth Information
Slicing and dicing our data
There are many interesting questions we can ask about our analytics observations; most of
them can't efficiently be answered by the
status_update_views
table. Here are a
few:
• How many total views did status updates receive on each day of September?
• What percentage of status update views are via the web, via our mobile app, and
via third-party applications?
• Which hours of the day are most popular for viewing status updates?
• What is the average monthly growth of status update views this year?
In order to answer any of these questions efficiently, we will require a precomputed table
that's structured with that question in mind. Instead of reading massive quantities of raw
observations into memory, and then computing aggregate information about it, we'll keep
the aggregate information up-to-date as we make the discrete observations.
