Using AI Highlights for Analysis

Alchemer Dashboard is currently waitlist only. Visit this page to learn more about Dashboard or join the waitlist!

Everywhere in Alchemer Dashboard where you can run AI Highlights analyze, you can configure a customized AI Highlights analysis. The customized analysis can be a one-time customization or can apply to all future analysis of that particular data source.

Finding customization controls

Every menu or location where you can run AI Highlights analyze, you can customize your analysis. You access using the more options menu. 


AI Highlights window workflow

1. The AI Highlights analysis window first asks you what kind of analysis you would like to do:

  • Outlier analysis: finds underlying anomalies or outliers
  • Trend analysis: finds interesting trends in the underlying data
  • Cross-correlation analysis: determines how your metrics correlate with other measures
  • Change analysis: if you select 2 or more data points, you can run a change analysis on those points. Otherwise, Alchemer Dashboard does not show you this option.
If you enable Forecasting, AI Highlights only allows you to run one kind of analysis at a time. If you disable Forecasting, you will be able to multi-select.
2. Next, AI Highlights recommends certain columns for analysis. You can select all columns in the Chart, clear all columns, or search for more columns to include.

3.  You can also customize advanced parameters. Select Customize parameters, on the Select measures and attributes page. The following screen appears. Scroll down to see more options, including customizations for specific analysis types.

You can customize the following parameters for any AI Highlights analysis except a change analysis:

  • General settings: exclude null values, restrict analysis to the current result set only, exclude zero values, or auto-tune date boundaries
  • Insight count settings: choose the maximum number of measure and attribute columns to analyze

You can customize the following parameters for an outlier analysis:

  • Outlier algorithm: choose the method of outlier detection (z-scores or median z-scores, Seasonal Hybrid ESD, or Linear Regression)
  • Outlier parameters: choose the minimum rows required for analysis, multiplier for outlier detection, and maximum P-Value
  • Insight count settings: choose the maximum outlier detection insights, maximum seasonality detection insights, and maximum linear regression insights

You can customize the following parameters for a trend analysis:

  • Trend parameters: specify the minimum rows required for analysis, maximum P-Value, and minimum relative difference
  • Insight count settings: specify the maximum trend insights

You can customize the following parameters for a cross-correlation analysis:

  • Cross-correlation parameters: choose the minimum rows required for analysis, minimum correlation coefficient, maximum correlation coefficient, and maximum correlation lag
  • Insight count settings: choose the maximum cross-correlation insights

You can customize the following parameters for a change analysis:

  • Change analysis parameters: specify the maximum difference elements, the maximum fraction, minimum abs change ratio, and minimum change ratio

You can also reset these parameters to their system defaults. Select Reset to system default in the lower-left corner of the AI Highlights modal.

4. Select Analyze.

Basic Standard Market Research HR Professional Full Access Reporting
Free Individual Team & Enterprise
Feature Included In