Compare data across time periods, including growth and relative analysis, without writing a formula.
A time series is a set of data points ordered by time — for example, a sale in January 2020 comes before a sale in February 2020. Use Invoca’s time series analysis feature to search for Dashboard Tiles about series data, such as comparing sales for each month across several years, calculating growth over the same time period across other periods, or relative analysis like sales for the last 3 months of each year across several years.
Period keywords
Examples: orders by day; sales day of month by month; deals day of quarter by year; customers by week day of week; ad clicks day of year by year; visitors by hour weekly; purchases month of quarter by year; purchases by quarter; opportunities quarter of year last 4 years yearly; product shipments week of month yearly; revenue week of quarter; new products week of year last 3 years yearly.
All of these keywords sort data using datetime semantics — chronologically in a time sequence. When you type one into the search bar, Invoca asks which data source to apply it to; by default the Search bar suggests these keywords less frequently than others.
You can combine these with the existing data keywords: Detailed, Hourly, Daily, Weekly, Monthly, Quarterly, Yearly.
Examples of time series analysis
Typically, when you search for Dashboard Tiles about series data, the visualizations that answer your questions are line charts, which frequently (but not always) include a stack to indicate a period.
A typical response to a time-series search is a line chart showing the aspect as it rises and falls over time. You can also add a relative date filter, for example:
The child date-time attribute is on the x-axis, and the parent appears in the legend. For example, searching revenue month yearly puts the child, monthly, on the x-axis and the parent, yearly, in the legend.
Granularity for date filters
Refine simple date filters by adding a hierarchical date filter that specifies two bucket granularities, such as “hour of day” or “week of year”. The syntax is:
INTEGER_CONDITION is optional but must be an integer. For example, revenue by day of week <= 2 is valid; revenue by day of week = Tuesday is not. You can specify one or more granular filters.
Tips and gotchas for time granularity:
- System-defined fiscal rules are respected — if the fiscal year begins in February,
month of year = 2 matches dates in March.
- Fiscal shorthands like
Q1, Q2 aren’t supported, so day of week = d1 isn’t valid.
INTEGER_CONDITION with = or != accepts a list of filter values, so day of week = 1 2 3 is valid.INTEGER_CONDITION with = requires legal values (1-7 for day of week), while > accepts any integer.- Simple date filters can be edited through the Dashboard Tile to refine your search; adding a hierarchical date filter in the search bar disables this ability.
Create a max(date) field and use it to filter
If you have a date field and want to return the most recent set of data based on it:
- Create a formula called
Max Date, for example:
- In the search bar, filter your dates by this formula, for example:
max date = true. This returns only fields that pass the filter.
Where to go next
Last modified on September 23, 2026
