StatResult Parser Pro

Extractor-only GPT that auto-detects statistical result tables and structures them into a common JSON schema with minimal loss.

Overview
Version
v1.0.0
Created
2025-12-14
Updated
2025-12-14
statisticsdata-extractiontable-parsingspssrpythonreproducible-research
statresult-parser-prostat-parserresult-extractor
Key functions
  • Auto-detect and split tables from Excel/CSV/TSV/HTML/Markdown/text outputs (e.g., SPSS console) and some well-structured PDFs
  • Assign table_ids (t001, t002, …) and convert each table into a shared JSON schema (dimensions + cells)
  • Identify analysis types (e.g., descriptive, crosstab, regression, anova, reliability, hayes, pls) using title/header patterns
  • Represent missing/symbol-only cells as null (or explicit strings) and record parsing rules/limitations in notes
Technical details
_id
g-689d215bc6bc8191be5a2886cd559f53
gpt_id
g-689d215bc6bc8191be5a2886cd559f53
viz1
public
viz2
show_url
language
en
Other fields
additional_features
["Heuristic-based analysis type detection (regression/ANOVA/reliability/crosstab/Hayes/PLS, etc.)"]
example_commands
["From the attached SPSS output (.txt), detect every table and convert them into the provided JSON schema. Preserve titles/section names when available.", "Parse all sheets in this Excel file: treat each non-empty cell block separated by blank rows/columns as a separate table and output JSON with t001… ids.", "Convert the three Markdown tables below into dimensions/cells JSON. Treat '-' and 'n/a' as null and document the rule in notes."]
gpt_id
g-689d215bc6bc8191be5a2886cd559f53
ideal_use_cases
["Batch-extract SPSS-style output tables (ASCII) such as Descriptives/Correlations/ANOVA/Coefficients/Reliability into JSON", "Detect multiple table blocks per Excel sheet (separated by empty rows/columns) and enumerate them as t001…", "Convert HTML/Markdown tables into structured JSON for downstream ETL or database loading", "Preserve subtype cues for Hayes PROCESS / PLS-style outputs when keywords are present"]
limitations
["Does not interpret results (no hypothesis decisions, no causal claims, no inferred variable roles); extraction only.", "PDF parsing is attempted only when table structure is clear; otherwise limitations are documented in notes.", "Highly complex merged cells/multi-level headers/footnote layouts may reduce label normalization fidelity (original labels are preserved when possible)."]
target_users
["Researchers/graduate students who need reusable structured outputs from SPSS/R/Python/PLS result tables", "Analysts moving statistical tables into pipelines (JSON/DB) for automation and reproducibility", "Users who want automatic table block detection/normalization across multiple sheets/files"]