Alchemer Dashboard is currently waitlist only. Visit this page to learn more about Dashboard or join the waitlist!
The AI Highlights Analyze feature works without you having to do anything but push a button. However, like any other feature, there are things you can do to optimize the feature. This page contains some best practices you can use to make AI Highlights more effective when you use it.
When to invoke AI Highlights Analyze
A good time to run AI Highlights Analyze is right after you upload data. AI Highlights Analyze can very quickly help you find insights in your data.
Start from a Search. Enter a single measure in the bar; one you want to explore, of course! Then, select the More menu > AI Highlights analyze. Choosing the single measure focuses AI Highlights.
Customize your analysis to focus or tweak the AI Highlights results. While you are tempted to keep all the columns, eliminating some can also result in a better analysis.
Do your data modeling
Increase AI Highlights’s effectiveness by ensuring you are practicing good data modeling. This is true if you are a user uploading the occasional data file or a data management professional. Modeling data requires that you can:
1. Select Data to get to the data management listing.
2. Select a data source you own or can edit. This brings up the Columns screen, where you can make your modeling settings.
3. Modify one or more column settings.
4. Save your changes.
Make sure you set the INDEX PRIORITY for columns in your data source. Use a value between 8-10 for important columns to improve their search ranking. Use 1-3 for low priority columns. INDEX PRIORITY impacts user-based ranking which helps AI Highlights focus its analysis.
AI Highlights Analyze uses measures for correlations. For trendlines and outliers, if AI Highlights Analyze has a measure, it then drills by attributes in turn.
ATTRIBUTE = text or date that you cannot sum
MEASURE = values you can math on, with a meaningful result
Attributes
- Fruit
- Grocery
- Macintosh
Measures
- Price
- Age
- Weight
What about?
A style number or product ID
You should also set AGGREGATION on your columns. AI Highlights Analyze applies the default aggregations from your data when it pulls measures for analysis.
Situations to avoid
There are some use cases AI Highlights is not yet designed to handle.
If your data contains a measure that uses a
MOVING_*orGROUP_*formula, AI Highlights may return results that simply aren’t meaningful.When doing a correlation analysis, AI Highlights may not find meaningful data if you have a measure with anything other than
SUM.
Set AI Highlights preferences
To set preferences for AI Highlights, click AI Highlights in the top bar, and then select Default preferences. These preferences allow you to control how you receive analysis notifications. They also allow you to exclude nulls or zero value measures from analysis.
The exclusions impact each AI highlight. It eliminates points with such values during statistical calculations for example, for mean, standard deviation AI excludes values from any equation and uses only the remaining points.
For more information, refer to AI Highlights preferences.
Prioritizing analyses types
You can prioritize highlighting changes in data over time instead of other changes, such as outliers or anomalies.
Columns
When you trigger a AI Highlights analysis on a Chart, you can select alternate data columns. To trigger more time-related insights, pick more date-time columns.
Advanced
In the advanced tab of the AI Highlights dialog, increase the maximum number of trend and correlation insights, and reduce the number of anomaly insights.
