Analysis & Training

The Analysis & Training page provides an interactive dashboard to help users better understand what's hidden behind the issues’ data and train the application on a specific data selection.

Data gathering

There is no need to do anything: data extraction is fully automated 🤖
Nostradamus starts data gathering as soon as the application is started.

Filtering

You can easily filter your data by specific conditions using the Filter card. Initially, it has all the basic Jira-issue fields, but if you need to add some other fields and/or edit the existing ones, they can be configured in the Settings menu.

The application provides the following filtering types:

• Text Search can be performed using one of the following options:

1. full-match (the exact-match flag is switched on)
2. substring search (the exact-match flag is switched off)

• Drop-down list

1. full-match (the exact-match flag is switched on)
2. subset search (the exact-match flag is switched off)

• Numerical range
• Date range

Defect submission chart

This chart represents the dynamics of defect submission for a chosen period.
The following periods are available to choose from:

• Day
• Week
• Month
• 3 Months
• 6 Months
• Year

Frequently used terms

The card presents the terms which are frequently used in bug descriptions.

Significant terms

The card presents the most significant terms – based on significance weight – that are extracted from bug descriptions. The terms can be calculated for the following categories which are based on your data:

• Priorities
• Resolutions
• Areas of testing (have to be added on the Settings page)

Statistics

Statistical info for the following metrics:

• Attachments (the number of attached files)
• Comments (the number of comments)
• Time to Resolve

Training

The application can be trained using the filtered data to predict different metrics based on bug attributes. To start training, just configure the training settings in the Settings menu and hit the Train Models button.