Skip to main content

Interface Tour

This page describes the main areas of the DashAI interface and what each one does.

The top bar is always visible and provides access to all main modules:

SectionWhat it does
DATASETSLoad datasets, explore their content, and open notebooks
MODELSCreate sessions, train models, compare results, generate predictions
GENERATIVEInteract with text and image generation models
PLUGINSInstall and manage plugins

The language dropdown (next to the globe icon) switches the interface language (between english and spanish). The hardware monitor button displays system resource usage. The tutorials button opens guided walkthroughs. The theme button toggles between light and dark mode.

The sidebar content changes depending on which module you are in.

In Datasets: Shows Available Datasets (with row and column counts) and Notebooks (grouped under their parent dataset). A search bar filters both. The New Dataset/Notebook button at the top starts the upload or notebook creation flow.

In Models: Shows Available Datasets and Sessions grouped by task type. A search bar filters datasets and sessions. The New Session button creates a new session.

In Generative: Shows sessions organized by generative task type.

Main Area​

The central content area changes based on what you have selected.

Datasets — Dataset View​

When you click a dataset in the sidebar, the main area shows the dataset's EDA panel:

  • Header — dataset name, creation date, Quality Score, and NEW NOTEBOOK button
  • Summary cards — Total Rows, Total Columns, File Size, Duplicated Rows, Missing Values
  • Quality banner — green checkmark if no issues, warning if issues were detected
  • Analysis tabs — Overview, Numerical Analysis, Categorical, Text, Data Quality, Correlations

Datasets — Notebook View​

When a notebook is open:

  • Notebook title — Notebook: [Dataset Name] Preview
  • Toolbar — FILTERS and EXPORT controls
  • Dataset preview table — paginated view of the data in its current state; updates after each converter
  • Operation timeline — each Explorer or Converter you add appears below the preview as a block with its result, status badge, and edit/delete controls
  • SAVE AS NEW DATASET button — top right, saves the current state as a new dataset

Models — Session View​

When a session is open:

  • Model Comparison panel — table of all models with their metrics; toggle between TRAINING, VALIDATION, and TEST splits; switch between TABLE and CHARTS views
  • Model cards — one per model, expandable, each with EDIT, TRAIN/RE-TRAIN buttons, a status badge, and four inner tabs: LIVE METRICS, EXPLAINABILITY, PREDICTIONS, HYPERPARAMETERS

Generative — Chat View​

When a generative session is open, the main area shows a chat interface:

  • Chat history — a scrollable conversation log showing your messages and the model's responses
  • Message input — text field at the bottom to type your prompt; press Enter or click the send button to submit
  • Model responses — displayed inline in the conversation, with support for formatted text

Right Panel​

The right panel is context-sensitive and appears alongside the main area.

In Notebooks (EXPLORE tab): The Analysis Tools panel lists all available explorer tools organized by category. Hover over any tool to see a preview image and description. Click to open the configuration modal.

In Notebooks (CONVERT tab): Same panel, but showing converter tools organized by category.

In Models: The Available Models panel lists all models compatible with the current session's task. Hover for a description. Click to open the Add Model modal.

In Generative: Shows the hyperparameters of the selected model, allowing you to modify them directly before or between interactions.

Configuration Modals​

Many tools open a two-step configuration modal:

Step 1 — Configure Scope: Select which columns the tool will use. The column selector table shows index, name, value type, and data type. A counter at the top shows how many columns are selected and how many are required.

For converters, Step 1 also includes row scope — select rows by range, by specific indices, or use SELECT ALL.

Step 2 — Configure Parameters: Each tool has its own set of parameters, auto-generated from the tool's schema. Each parameter has a ? help icon. Click CREATE EXPLORER or CREATE CONVERTER to apply.

Job Queue​

The Job Queue is visible at the bottom right of the screen whenever jobs are running or have recently completed. It shows:

  • Active jobs (with a spinner and count)
  • Failed jobs (with a red indicator and count)
  • A Show Completed button to review finished jobs

Long-running operations — training, exploration, prediction generation — all run as background jobs so you can continue working while they process.

Status Badges​

Status badges appear on model cards and explorer/converter results throughout the interface:

BadgeMeaning
Not StartedThe operation has been configured but not yet run
FinishedThe operation completed successfully
ErrorThe operation failed — check parameters or data
StartedThe operation is currently running
DeliveredThe operation has been queued