Data Collection

Upload Datasets

Select method

CSV Excel Ingest SQL Query Ingest Full Table

CSV

Delimiter Leave blank to auto-detect

Excel

Sheet Name

Ingest SQL Query

Write a custom query

Connection Query Dataset name (optional)

Ingest Full Table

Connect & copy a table

Connection

Table Row limit

Dataset name (optional)

Datasets

{{d.name || d.id}} {{d.id}} ({{d.rows}}×{{d.cols}})
No datasets yet. Upload one above.

Preview

Filter column: Op: Value:

Sort by:

Rows {{vm.offset}} - {{vm.offset + vm.limit}} of {{vm.total}}

{{c}}
{{r[c]}}

Integration

Validation: {{ vm.validation.ok ? 'OK' : 'Has issues' }} {{ vm.validation.warnings.length }} warning(s) {{ vm.validation.errors.length }} error(s)

Errors

{{e}}

Warnings

{{w}}

Semantic Model

Model name

Primary dataset

Available datasets

{{d.name || d.id}} in model

Model sources

{{sid}} {{ (vm.dsColumns[sid] && vm.dsColumns[sid].length) ? (vm.dsColumns[sid].length + ' cols') : 'loading cols' }}

Relationships

Left dataset
Column

Right dataset
Column

Join type

{{r.leftDs}}.{{r.leftCol}} {{r.type || 'inner'}} join {{r.rightDs}}.{{r.rightCol}}

Advanced (optional)

Use Dimensions to group by fields, and Measures to compute aggregations like sum, avg, count distinct. You can ignore this section and just Preview/Export the integrated join.

Dimensions

Source Column Alias

{{d.source}}.{{d.column}} as {{d.alias}}

Measures

Source Column Aggregation Alias

{{m.agg || 'sum'}}({{m.source}}.{{m.column}}) as {{m.alias}}

Saved Models

{{m.name}} ({{m.id}})

Integration Summary

Primary: {{ vm.integrationSummary.primary || '-' }}

Sources: {{ vm.integrationSummary.sourcesCount }}

Rows: {{ vm.integrationSummary.rows }} Columns: {{ vm.integrationSummary.columns }}

Sample columns:

Relationships

{{rel}}

Summary reflects the latest Preview result.

Integrated Preview

Rows {{vm.semanticPreview.offset}} - {{vm.semanticPreview.offset + vm.semanticPreview.limit}} of {{vm.semanticPreview.total}}

{{c}}
{{r[c]}}

AI Assistant: Semantic Model Builder

Provide a goal; the assistant will suggest dimensions, measures, and relationships based on current data.

Goal / Prompt

Context (auto):

{{ vm.aiContext() }}

Apply to model

Existing semantic model New model name

Tip: choose an existing model to merge suggestions into it, or leave it blank and provide a new name.

Source Datasets (required for schema visualization)

No datasets available

Note: Select at least one dataset so the schema can be visualized. You can add more sources later in the Integration section.

Debug info:
Current origin: {{vm._origin || 'unknown'}}
API endpoint: {{vm._apiUrl || 'http://127.0.0.1:8000'}}

Suggestions

Rationale:

{{r}}

Dimensions

{{d.name}} — cols: {{d.columns.join(', ')}}

Measures

{{m.name}} = {{m.expression}}

Relationships

{{rel.fromTable}}.{{rel.fromColumn}} → {{rel.toTable}}.{{rel.toColumn}} ({{rel.type||'inner'}})

Processing

Available DataFrames

Datasets ({{vm.datasets.length || 0}})

{{ds.name}} No datasets found. Upload CSV files in Data Collection section.

Semantic Models ({{vm.semanticModels.length || 0}})

{{m.name}} No semantic models found. Create one in Data Collection section.

Selected Model Tables ({{vm.proc.tables.length || 0}})

{{t.var}} Select a semantic model below to see its tables Click "Refresh tables" to load

Notebook Outputs ({{vm.proc.outputs.length || 0}})

{{name}} Run cells to create output dataframes

Notebook

Semantic model:

Notebook:

Name:

Cell {{$index+1}}

{{vm.proc.stdouts[$index]}}

Detected DataFrames

DataFrames created in cells (variables) appear here after running. Save them as datasets to reuse in Analytics and Visualization.

{{name}}

{{c}}
{{r[c]}}

Run your cells to see outputs here.

{{grp.name}} ({{grp.smId}}) {{grp.datasets.length}} table(s)

Integrated {{d.name || d.id}} {{d.rows}}×{{d.cols}}

No datasets yet. Upload/ingest one in Data Collection.

Selected

{{ (vm.datasets | filter:{id: vm.ml.datasetId})[0].name || vm.ml.datasetId }} {{ (vm.semanticModels | filter:{id: vm.ml.smId})[0].name || vm.ml.smId }} Integrated -

Pick a semantic model's Integrated chip or a specific table chip.

Columns ({{vm.selectedSchema.columns.length}})

Name	Type
{{c}}	{{vm.selectedSchema.dtypes[c] \|\| '-'}}

Rows: {{vm.selectedSchema.rows || (vm.selectedSchema.total) || 'n/a'}}

Sample (first 5 rows)

{{c}}
{{r[c]}}

Method Recommendation

Select a dataset or semantic model (left panel) to enable recommendations.

{{i}}

Source: {{vm.recommend.summary.source_type}} {{vm.recommend.summary.source_id}} · Rows {{vm.recommend.summary.rows}} · Cols {{vm.recommend.summary.columns}} · Numeric {{vm.recommend.summary.numeric_columns}} · Categorical {{vm.recommend.summary.categorical_columns}} · Target {{vm.recommend.summary.target}} (distinct {{vm.recommend.summary.target_distinct}})

Recommended Methods: {{m}}

Reasons:

{{r}}

Issues:

{{i}}

Analyzing...

Method: Target: RL Method: Prediction Mode: ⚡ No training needed Episodes: API Key: Initial Balance: k: Features:

Accuracy: {{ (vm.ml.trainSummary.accuracy*100) | number:1 }}% · Precision: {{ (vm.ml.trainSummary.macro_precision*100) | number:1 }}% · Recall: {{ (vm.ml.trainSummary.macro_recall*100) | number:1 }}% · F1: {{ (vm.ml.trainSummary.macro_f1*100) | number:1 }}% {{vm.ml.trainCopyMsg}}

Hyperparameters

max_depth: min_samples_split: n_estimators: n_neighbors: hidden_layer_sizes: max_iter: C: sequence_length: epochs: batch_size: units: task: learning_rate: gamma (discount): epsilon (exploration): epsilon_decay:

Tips

Use Integrated to train on the joined model; or pick a single table.
For ANN, use hidden_layer_sizes like “64,32”.
For Random Forest, increase n_estimators for better accuracy.
For LSTM, ensure data is ordered chronologically; sequence_length defines past observations to use.
For RL, use stock data with OHLCV columns. Agent learns to buy/sell/hold to maximize returns. Q-Learning is simpler, DQN uses neural networks.

Reinforcement Learning Results

Training Summary

Method: {{vm.ml.result.method}}

Episodes: {{vm.ml.result.episodes}}

Initial Balance: ${{vm.ml.result.initial_balance | number:2}}

Best Portfolio: ${{vm.ml.result.best_portfolio_value | number:2}}

Return: {{vm.ml.result.final_return_pct | number:2}}%

Best Episode: {{vm.ml.result.best_episode}}

Last 20 Episodes Portfolio Values

Result

{{ vm.ml.result | json }}

Model ID: Target:

Accuracy: {{ (vm.ml.evalSummary.accuracy*100) | number:1 }}% · Precision: {{ (vm.ml.evalSummary.macro_precision*100) | number:1 }}% · Recall: {{ (vm.ml.evalSummary.macro_recall*100) | number:1 }}% · F1: {{ (vm.ml.evalSummary.macro_f1*100) | number:1 }}% · AUC: {{ vm.ml.evalSummary.auc | number:3 }} {{vm.ml.evalCopyMsg}}

MSE: {{ vm.ml.evalSummary.mse | number:4 }} · MAE: {{ vm.ml.evalSummary.mae | number:4 }} · R²: {{ vm.ml.evalSummary.r2 | number:4 }}

Inertia: {{ vm.ml.evalSummary.inertia | number:3 }}

Final Portfolio: ${{ vm.ml.evalSummary.final_portfolio | number:2 }} · Return: {{ vm.ml.evalSummary.total_return_pct | number:2 }}% · Total Trades: {{ vm.ml.evalSummary.total_trades }}

Confusion Matrix

ROC Curve (binary)

RL Agent Performance

Final Portfolio: ${{vm.ml.eval.final_portfolio | number:2}}

Total Return: {{vm.ml.eval.total_return_pct | number:2}}%

Total Trades: {{vm.ml.eval.total_trades}}

Buy Actions: {{vm.ml.eval.buy_count}}

Sell Actions: {{vm.ml.eval.sell_count}}

Hold Actions: {{vm.ml.eval.hold_count}}

Portfolio Value Over Time

Regression Metrics

MSE: {{vm.ml.eval.mse}}
MAE: {{vm.ml.eval.mae}}
R²: {{vm.ml.eval.r2}}

Clustering Metrics

Inertia: {{vm.ml.eval.inertia}}

Cluster Sizes

Model ID: Output dataset name: Prediction column: include probabilities

AI Assistant: ML Pipeline Builder

Describe the goal; the assistant will suggest a model, target, features, and parameters based on the current selection.

Goal / Prompt

Suggested Pipeline Switch to: {{vm.analysisAi.pipeline.suggestedDatasetId}}

Task: {{vm.analysisAi.pipeline.task}} · Method: {{vm.analysisAi.pipeline.method}} · Target: {{vm.analysisAi.pipeline.target}}

Features:

{{f}} -

Params:

{{k}}: {{v}}
(none)

Rationale:

{{r}}

Visualization

Build & Save

Source: Dataset: Semantic Model: Pick a source to load columns.

Chart: X: Y (one or more):

Name: ID:

Saved visualizations

No visualizations yet.

{{v.name}} ({{v.viz_id}} · {{v.source_type}}: {{v.source_id}})

Charts

Quick exploratory charts for your data source. Select a source above in "Build & Save" section.

Select a data source in the "Build & Save" section above to enable charts.

📊 Histogram

Column: Bins:

📈 Value Counts (Top Values)

Column: Top:

🔥 Correlation Heatmap

Shows correlations between numeric columns.

⚫ Scatter Plot

X: Y:

📦 Boxplot (Distribution)

Column:

Settings

Layout

Bookmarks

Edit Profile

Data Collection

Upload Datasets

CSV

Excel

Ingest SQL Query

Ingest Full Table

Datasets Refresh

Preview {{ vm.previewCollapsed ? 'Expand' : 'Collapse' }}

Integration

Errors

Warnings

Semantic Model

Available datasets

Model sources

Relationships

Advanced (optional)

Dimensions

Measures

Saved Models

Integration Summary

Relationships

Integrated Preview

AI Assistant: Semantic Model Builder

Suggestions

Dimensions

Measures

Relationships

Processing

Available DataFrames

Notebook

Detected DataFrames

Analysis

Datasets for Analysis

Method Recommendation Analyze

Train Model

Reinforcement Learning Results

Result

Evaluate Model

Confusion Matrix

ROC Curve (binary)

RL Agent Performance

Regression Metrics

Clustering Metrics

Cluster Sizes

Predict & Save

AI Assistant: ML Pipeline Builder

Suggested Pipeline Switch to: {{vm.analysisAi.pipeline.suggestedDatasetId}}

Visualization

Build & Save

Saved visualizations

Charts

Create New Project

Share Project

Add User

Users with Access

Datasets

Preview

Method Recommendation