Local Upload
Import & Merge
Import the current data first, then configure how the second dataset is merged, so the entries don't crowd into one block.
Import Current Data
Local upload, platform database, and paste-text are split into separate cards.
Database
Search first, then select a table
Importing from the platform database loads the entire table. To reduce variables, use "Delete Variables" after importing.
No table selected
Paste Text
Merge a Second Dataset
Data source, merge method, and execution are split into three blocks for a more stable layout.
Second Data Source
No table selected
Selecting a second dataset from the platform database also imports the entire table.
Merge Method & Keys
Run Merge
Append aligns columns automatically; horizontal bind requires equal row counts; key matching reports match rate, duplicate-key risk, and unmatched rows.
Variable Operations
Rename, label, type conversion, and deletion are handled separately—one action at a time.
Rename Variables
Rename one by one to avoid crowded, error-prone bulk editing.
Import data to batch-edit variable names.
Variable Labels
Add academic meaning or source notes for individual variables, useful for export and paper writing.
Import data to batch-edit variable labels.
Variable Type Conversion
Suitable for batch conversion; the result reports the number of failed conversions.
Supports %td: Stata daily date counts since 1960-01-01, text such as 01jan1960 or 1960-01-01, and already parsed dates. For other date text, use a format such as %Y-%m-%d.
Delete Variables
Deletion is a separate card, kept apart from renaming and type conversion.
Deletion creates a new version; you can still restore via undo or version history.
Sample Operations
Missing values, duplicates, outliers, filtering, and sampling are split into separate cards with aligned parameter grids.
Missing Value Handling
Duplicate Handling
Leave empty to check duplicates across all variables.
Outlier Handling
Random Sampling
Conditional Filtering
Conditional filtering is now a separate large card to avoid being crowded by other features.
The filter supports limited-depth AND/OR groups; the backend accepts only structured JSON, not arbitrary code expressions.
Variable Generation
Each generation method is a separate card to avoid crowding too many list boxes and inputs on one screen.
Questionnaire scoring and coding
Complete-item totals require every item. Means and prorated totals use the minimum valid item count. Repeat the operation for additional scales.
New variables preserve the original items. Reverse scores equal the sum of the scale limits minus the original score. Multiple-response coding uses one column per slot; unanswered rows remain missing.
Log Variables
Ratio Variables
Interaction Terms
Supports 2 to 3 numeric variables; for 0/1 dummies, the result note explains the interaction meaning.
Dummy Variables
Lag Term
Only lags need panel structure settings; if the time variable has duplicates without an entity variable, generation is blocked.
Difference Term
Only differences need panel structure settings; if the time variable has duplicates without an entity variable, generation is blocked.
Standardization
Centering
Composite Index
The entropy method, PCA, and factor analysis are split into three independent cards for easier configuration and interpretation.
Entropy Method
Direction settings are generated after selecting indicators
PCA (Principal Component Analysis)
Factor Analysis
Current Data Snapshot
The previously empty lower section now shows a summary that updates with the data state.
Current Source
Not imported
Data Size
0 rows / 0 columns
Version Progress
v0 / v0
History
0 records
Import data first; the system will then create undoable processing snapshots.
Current Feature Guide
Switching features on the left updates the input and output hints here.
The workspace keeps only the main action card for the selected feature; other instructions are collapsed here.