Aggregate data by date

The Research report includes several options for examining the data in your presentation by units of time. For example, if you have a dataset that shows what days and times people visit certain websites, you could create a presentation breaking out the data by day of the week over the course of a month to see if certain days have higher usage than other days, and whether those days might correlate with other activities (online promotions or coupons, holidays, etc.).

In this article, we'll show you haw to use date-based aggregation categories. There are three parts necessary to configure a presentation this way:

  • a dataset with time-based data
  • carefully designed data series that focus on a specific date range
  • a grouping that is time-based


  1. Start a new presentation.
  2. Select the dataset you want to use. Make sure it has time-based data.

    It could be any type of information - like a number of pizzas eaten per month, pet supply sales volume per week, TV viewers per half hour, or vehicle traffic per minute.

    Data that you upload yourself can be used with this system, as long as it has a calendar-based time element (year, quarter, month, week, day, hour, or minute).

    This feature is still under development; not all syndicated datasets with temporal elements are currently supported. If you encounter problems, contact Rhiza by using the Feedback and Support link.

    Rhizalytics can aggregate data by the smallest unit in the dataset, or by larger units. That is, if your dataset reports sales by month, you can show months, quarters, and years in your report, but you can't show sales by week or day because the data isn't available at that level of precision.

  3. Create at least one data series with a date filter to specify the overall time period to show. In many cases, you'll want to set up multiple series with a consistent date range.
    You can set different date ranges in different data series, but remember that the grouping will not include all data series if the time periods don't overlap. For example, if you break out by month, a data series that covers a full calendar year will be represented in the January category, but a data series that covers only the second quarter of that year will be missing from the January grouping. Different date ranges in data series also affect index calculations.
    1. In the Data Series section, Click Filter Data Series to open the Edit Series window.
    2. In the Series Name field, type a meaningful name to represent the series.
    3. Use the Date filter to specify the time period you want to include. The Date filter
    4. In the Choose Filters section, add the first filter or you want to use.
    5. Add one or more values for the attribute.
    6. Add additional attributes and values until your series filters the data according to what you want to see, and then click Save.
    7. Repeat this process for any additional series you want to create.
  4. Use the Groupings section to choose how to show your data by time. In the list, the temporal categories are labeled at Date by time_element.

    Temporal groupings

    There are two options for each temporal grouping: the individual dates (for example, Date by Month) and the group date (for example, Date by Month (Group)). The group options collect data for all similar time periods in one grouping - that is, Date by Month (Group) aggregates data for all January months together.

    These screen shots show the same report with different aggregation groupings. The data series include the CYTD (12 months of data).

    Aggregating to a date By Month (Group) grouping

    Aggregating to Date By Month grouping

    Remember that if you choose a grouped grouping, the number of months (or days or quarters) included might be different from one group to another. The example above shows 12 months of auto sales from January 2015 through December 2015.

    If it showed data from a longer time period -- let's say from January 2014 through March 2015 -- the report would include two instances each of January, February, and March, but only one each for April, May, June, July, August, September, October, November, and December. If the chart showed the total number of registrations in January instead of the index values, you would expect the months with two years' worth of data to show about twice as many sales as the months with one year's data.

  5. Click Apply Changes.

© 2016, Rhiza, Inc. All rights reserved. Last updated March 02, 2016 01:36:41 PM.

Legal Information