You are here: Create presentations > Examine data in your presentation

Examine data in your presentation

Rhiza for Research offers different methods for digging into the data:

  • Visualizations
  • Calculated values
  • Context series
  • Aggregation by time-based, map-based, or dataset-specific categories

We encourage you to explore your information with a combination of these tools and techniques. Below are a few tips to get you started.

Tip 1: Try different visualizations to look at the target series

For an overview of visualization types, see Visualizations

Tip 2: Use a combination of visualizations to tell the complete story

Sometimes, a combination of a table and a map works better than either one alone. Or side-by-side maps highlighting different data points can help illustrate your point. Our point: Play with the visualizations and see whether a combination of them tell a better story than a single one might.

Tip 3: Instead of using an attribute in a filter, try aggregating to it as a Grouping (or vice versa)

Get creative in how you approach the data. Maybe you think you want to create data series for different pet types and then group them by their owners' political party. Let's see what that might look like:

Looking at pet ownership by political party, using political parties as the grouping

There's a lot of noise in there that we don't want, and it's coming from the grouping; there are so many responses to the question of party description. So, let's flip this around and use the party description to form our data series and the type of pet owned as the grouping.

Using political parties as data series rather than as a grouping

Our result might look something like this:

The result of reversing a grouping a data series

Tip 4: Pay attention to the sample size of the data

When you're looking at data from a survey-style dataset (for example, Scarborough USA+or L2 Nationwide Voter File), keep in mind that the companies who collect and maintain these datasets use modeling techniques to extrapolate the number of responses they get to reflect the trends in a larger geographic area. We recommend that you use the weighted individual and weighted household values, but that you also know take a peek at the raw number of respondents to determine how valid the results are. The more respondents there are, the more valid the weighted data.

Each data provider has its own threshold for what it considers unstable data; for example, Nielsen Scarborough considers anything with a respondent count lower than 70 to be "relatively unstable" and anything with fewer than 350 respondents to be "unreliable." If you're unsure of the validity of the data for a particular dataset, contact your dataset provider.

Tip 5: Use a context series to examine your data in light of a larger entity

Rhiza for Research lets you designate one data series as the context series so you can examine your data in light of a bigger market. See Create an Index with a context series for more information.

Tip 6: Understand calculated values in your visualizations

Visualizations offer different types of calculated values. See Calculated Values.

Tip 7: Know the nuances of your Dataset

How is the data collected? Is it weighted? What attributes are available? Are there any unexpected or unusual results that you might see when visualizing the data? To answer these questions, start with Using IHS Polk datasets: FAQs or Nielsen Scarborough FAQs. If you don't find answers there, refer to the dataset provider's website or your dataset provider customer representative.

© 2016, Rhiza, Inc. All rights reserved. Last updated April 06, 2017 04:04:44 PM.

Legal Information