Replies posted by haleemur
We have a calendar table in our Redshift instance, and we use it for the reasons you state above (holidays, weekend, marketing calendar events, etc.) If your organization maintains a Google Calendar, you can pull out events through their API, and store them in your calendar table. Then 3 years later, if an analyst wants to know what caused a spike, she / he can pull up the calendar annotations right next to the spike. Some time-series visualizations benefit from showing 0 values that would otherwise disappear from a regular GROUP BY on the table (e.g. line chart showing hourly sales for last 48 hours). To build such visualizations, it is necessary to left join from the date-hour table(s) to the data on the time field and aggregate up to the desired granularity, then the measures for periods with missing data will be converted to 0, and the can be rolled up as long as they’re additive or composed of additive measures (i.e. sum, average, min, max, etc…). median is an example of a non-add
Already have an account? Login
Login to the community
No account yet? Create an account
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.