Insight into your data
Kinds of Insights to expect from Data Assistant
Once you query the Data Assistant, you are presented with unique insights based on your data. These insights are delivered in a variety of formats to suit different analytical needs. Depending on the query.
The Data Assistant doesn’t just provide raw data in response to queries; the system highlights key insights from each query and offer recommendations for improving your data.
Visualisations
Data Assistant can generate different types of visualizations, including:
Bar Charts: Compare different categories and measure changes over time.
Line Graphs: Track trends and progressions to help with forecasting and pattern recognition.
Pie Charts: Understand proportional data and make easy comparisons between parts of a whole.
Area Charts: Used to track trends over time but with the added benefit of showing the magnitude of change by filling in the area beneath the line. Area charts are especially useful when showing volume or quantity changes and how they contribute to the overall trend.
When a Query is made:
You can get additional insight by clicking on the icons that appear below the responses
Download report as a pdf
Text to speech - An automated reading of the report
This views the sources of the data
This shows the metrics and dimension the Data Assistant is using
Sample:
The sources the data comes from
The Metrics and dimension the data Assistant is using to respond to the Query
The data model
The Skill, the data Assistant is using to respond to query.
Here’s a clearer breakdown based on the images above:
Data Source:
The system shows the origin of the data being analyzed. In this case, it is coming from Adobe Analytics (specifically labeled as "Adobeanalytics"), which tracks website traffic, user behavior, and other metrics related to digital analytics.
Metrics and Dimensions:
The metrics and dimensions that the data assistant uses to respond to the query are explicitly listed:
Metric: The specific measurement being analyzed (in this case, visits).
Dimensions: The data is segmented using different dimensions, such as:
Time breakdown by quarter of the year.
Different marketing channels.
Daily date ranges.
A custom eVar (eVar39) that tracks yearly data.
Data Model:
Two models are involved in the query:
Gpt-4o-mini: A language model that helps interpret and respond to the query.
DLM-v2: Likely a specialized data model used to handle and process analytics data.
Skill:
The system uses a specific "Analytics Skill" that applies the necessary algorithms, functions, or expertise to analyze the data. This skill ensures the right methods are used for data interpretation and response generation.
By providing insights into these aspects, the system makes it transparent how it processes and interprets queries, giving users a better understanding of the underlying mechanisms.
Last updated