Enhancing GoodData’s Semantic Layer With Natural Language

At GoodData, our mission is to empower companies to derive insights from their data via our API-first cloud-based BI platform. Providing users with the flexibility to enhance an ‘out-of-the-box’ solution and take advantage of our API-first headless BI architecture creates an exceptional user experience. By leveraging our technology partners’ third-party intelligent solutions and APIs, users can expect even more robust and customizable analytics that gain them a competitive advantage.

Our recently announced partnership with Arria NLG combines GoodData’s comprehensive data analytics and Arria’s natural language generation (NLG) technology to offer companies unparalleled insights into their data. This collaboration is aligned with GoodData's vision of providing organizations with a comprehensive analytics platform that goes beyond just collecting and analyzing data.

How Our New Partnership With Arria Will Benefit You

Integrating Arria NLG into GoodData's platform provides several enhancements that allow users to derive more value from their data. Some of the benefits include:

Automated Report Generation

Arria NLG allows for the automated generation of reports based on data insights. This means that businesses no longer have to spend time manually analyzing data and creating reports. With the click of a button, Arria NLG can generate reports highlighting key trends and insights.

Natural Language Generation

Arria NLG generates reports in natural language, making them easy to read and understand. This is particularly beneficial for users who may not have a technical background and struggle to interpret complex data. With Arria NLG, insights are presented in a way everyone can understand, reducing the risk of misinterpretation.

Report Customization

The integration of Arria NLG into GoodData's platform allows for the customization of reports based on user requirements. This means that users can tailor reports to their specific needs and preferences, ensuring they receive relevant insights into their business.

Improved Decision-Making

By providing users with meaningful insights in a format that is easy to understand, Arria NLG improves decision-making processes. Users can quickly identify trends and anomalies, enabling them to make informed decisions based on data insights.

How Are We Leveraging Arria’s Technology?

We built upon GoodData.UI library (a TypeScript framework for building analytical web applications on top of GoodData Cloud, GoodData Cloud Native, and GoodData Platform). The Arria technology — Arria Connect SDK (a software development kit used to make API requests) fits perfectly, as shown in the below image.

The only change here was to divide GoodData’s artifacts (Attributes, Facts, and Metrics) into Arria’s recognized data field types (dimensions and measures).

How the Natural Language Generation Process Works

The response from Arria’s service is in JSON format, and the tag that holds the actual narrative is encoded in HTML. Thus it can be immediately injected into GoodData’s UI interface and displayed (for example, beneath the chart).

Arria Connect SDK offers a set of pre-defined narratives that can be divided into four main categories:

1. Descriptive types
2. Variance types
3. Further analysis and Detection
4. Conversational mode

Let’s look at each of these categories in turn.

Descriptive narratives

The focus is on summarizing and describing the data clearly and concisely.

In the case of Descriptive statistics, this involves using the main statistical measures (summary, average, distribution, and data skew) to describe the characteristics of a dataset. It also tries to identify a leader’s category from insight data.

Arria also offers extensive descriptions of Bar charts, Line charts, and Pie charts.

Variance narratives

Variance is a statistical measure that indicates the spread or dispersion of a set of data points around their mean or average value. Time-based variance refers to the variance of a set of data points over time. This can be useful in identifying patterns or trends in data, such as seasonal fluctuations or changes in trends over time.

Further analysis and detection of narratives

Trend analysis is a statistical technique used to identify patterns and trends in data over time. It involves analyzing historical data and identifying patterns or changes that occur over a specific period.

Arria Connect SDK can perform a Trend analysis (changes in a measure over time with a significant shift highlight) and Ranking analysis (a summary of the top and bottom rankings of combinations of dimensions over time).

Furthermore, we can also use a detection narrative, specifically Anomalies detection (identification of data points or ranges that significantly deviate from statistical norms) and Correlations detection (generates an analysis of correlations between measures).

Conversational mode

Arria Connect SDK also allows a narrative called “Answers,” which enables you to have conversations with your data. The result? Your dashboard users can access instant insights and deeper detail using natural language.

Final Thoughts on This Exciting Integration

Integrating Arria NLG into the GoodData platform provides significant enhancements that will allow businesses to derive more value from their data. By automating report generation, providing natural language generation, allowing for customization, and improving decision-making processes, this new integration can further empower you to make data-driven decisions with confidence.

Ready to try GoodData? Sign up for our free 30-day trial today or to take a closer look at this integration, register for a demo.