restaurant-survey

Overview

This skill provides comprehensive analysis and conversational querying capabilities for the 2024 Restaurant Survey Dataset (file: 2409899.xlsx). The dataset contains detailed survey responses from 7,916 restaurant participants across 14,739 data fields, with complete data mapping and value encoding information.

Dataset Summary

File Information

• Filename: 2409899.xlsx
• Location: /c/Users/aayus/grey-chain/projects/runner_gpt/
• Total Sheets: 2
  • Sheet1 (A1): Survey response data
  • Sheet2 (Datamap): Complete data dictionary and encoding reference

Data Dimensions

Metric	Value
Survey Records (Respondents)	7,916
Survey Fields/Columns	14,739
Data Dictionary Entries	18,896 rows in Datamap

Sheet Structure & Analysis

Sheet 1: A1 - Survey Response Data

Characteristics

• Rows: 7,916 (header + 7,915 data rows)
• Columns: 14,739 fields
• Format: Each row represents one respondent's complete survey responses

Core Identifier Fields (Columns 1-8)

Column #	Field Name	Type	Description
1	record	Numeric	Record/respondent number (unique identifier)
2	uuid	Text	Participant universally unique identifier
3	markers	Text	Acquired markers (respondent attributes/segments)
4	status	Numeric (1-4)	Participant status (see encoding)
5	hCountry	Numeric (1-47)	Country from survey URL (see encoding)
6	qch	Numeric (1-5)	Channel information from URL (see encoding)
7	dTrack	Numeric (0-11)	Project version/change tracking
8	QSamp	Numeric (1-5)	Sampling type: Rep/Augment/Oversample (see encoding)

Data Field Categories

The remaining 14,731 columns contain survey question responses, organized into sections:

1. TimeSectionr1-r4 - Section timing data (start/end times in seconds)
2. Survey Questions - Main questionnaire responses (restaurant experience, preferences, etc.)
3. Hidden Questions - System-generated fields for analysis
4. Computed Fields - Derived metrics and calculated values

Sheet 2: Datamap - Data Dictionary & Encoding Reference

Structure

• Rows: 18,896
• Columns: 3
• Format: Hierarchical data dictionary with value mappings

Content Structure

The Datamap uses a hierarchical format with three columns:

Column	Purpose
Column A	Field description, variable name, or value label
Column B	Numeric code (for encoded values) or field references
Column C	Decoded text value or response option label

Key Data Encodings Found in Datamap

1. Status Field [status]

1 = Terminated
2 = Overquota
3 = Qualified
4 = Partial

2. Country Field [hCountry] - 47 Countries

1 = UK
2 = USA
3 = France
4 = Italy
5 = Germany
6 = Spain
7 = Sweden
8 = Denmark
9 = Norway
10 = Greece
... (37 more countries)
47 = Indonesia

Countries Included: UK, USA, France, Italy, Germany, Spain, Sweden, Denmark, Norway, Greece, Hungary, Czech Republic, Poland, Russia, South Korea, Thailand, UK (Welsh), Ireland, Australia, China, Japan, Austria, India, Turkey, UAE, Argentina, Brazil, Venezuela, Colombia, Switzerland, Portugal, Canada, Mexico, Malaysia, South Africa, Tunisia, Belgium, Netherlands, Israel, Saudi Arabia, Finland, New Zealand, Singapore, Chile, Taiwan, Hong Kong, Indonesia

3. Channel Field [qch]

1 = Open
2 = Loyalty
3 = Integrated
5 = 3PP (Third Party Partner)

4. Sampling Type [QSamp]

1 = Rep (Representative)
2 = Augment
3 = Oversample 1
4 = Oversample 2
5 = Oversample 3

5. Project Version [dTrack]

Tracks changes/versions from 0-11:

0 = No Changes
1-11 = Various updates (format changes, brand assignments, etc.)

6. TimeSectionr1-r4 (Timing Data)

Each section has three timing fields:

• c1 = Start time (in seconds)
• c2 = End time (in seconds)
• c99 = Duration/metadata

Data Patterns & Insights

Survey Structure

• Multi-section format: Questions organized into timed sections (Section r1-r4)
• Both open and closed-ended responses: Mix of numeric scales, text responses, and encoded values
• Respondent tracking: Full audit trail with timing, country, status, and channel information

Response Types Identified

1. Numeric responses: Restaurant ratings, frequency scales, agreement scales
2. Coded responses: Multiple-choice questions with numeric encoding (see Datamap for decoding)
3. Open text responses: Open-ended questions with text input
4. Time tracking: Duration data for each survey section

Geographic Coverage

47 distinct countries/regions represented in the data, with heavy representation from:

• Europe: UK, France, Germany, Spain, Italy, Nordic countries
• Asia-Pacific: Australia, Japan, South Korea, Thailand, China, India
• Americas: USA, Canada, Brazil, Mexico, Argentina
• Middle East & Africa: UAE, Saudi Arabia, Israel, South Africa, Tunisia

Sampling & Quality Flags

• Multiple sampling types: Representative, Augmented, and Oversample cohorts
• Status tracking: Terminated, Overquota, Qualified, and Partial responses
• Channel attribution: Responses sourced through Open, Loyalty, Integrated, or 3PP channels
• Version control: Project changes tracked through dTrack field (11 versions documented)

Data Query Capabilities

Questions You Can Answer

Respondent Filtering

• "How many respondents completed the survey?" (status = Qualified)
• "How many respondents are from each country?"
• "What's the breakdown by channel source?"
• "How many respondents are in each sampling group?"

Geographic Analysis

• "Which countries have the most respondents?"
• "Provide respondent counts by region"
• "Compare survey completion rates across countries"

Data Quality & Tracking

• "How many terminated, overquota, and partial responses are there?"
• "What changes were made in project version 3?"
• "How many respondents came from each survey channel?"
• "What's the distribution of respondents by sampling type?"

Timing Analysis

• "What's the average section duration for each part of the survey?"
• "Which section takes the longest to complete?"
• "Are there timing differences across countries?"

Response Analysis

• Survey questions can be filtered and analyzed using the encoded values from the Datamap
• Cross-tabulation of responses across demographics (country, channel, sampling type)
• Response distribution for Likert scale and multiple-choice questions

Example Queries

"Show me the survey response distribution across all 47 countries"
"How many qualified vs terminated respondents do we have?"
"What percentage of respondents are from the USA vs UK?"
"Display respondent counts by sampling type (Rep, Augment, Oversample)"
"Which survey section had the longest average completion time?"
"How many respondents came from Loyalty vs Open channels?"
"Show me the complete data map for understanding field encodings"

Data Mapping Reference

Column Groups

Administrative Fields (Columns 1-8)

• Record identification
• Participant attributes
• Survey metadata
• Sampling and quality information

Timed Sections (Columns 9-onward)

Multiple sections with timing data:

• TimeSectionr1: Pre-screener section timing
• TimeSectionr2: Main survey section 1 timing
• TimeSectionr3: Main survey section 2 timing
• TimeSectionr4: Demographic/follow-up section timing

Question Fields

Survey questions spanning all 14,739 columns, including:

• Restaurant experience questions
• Brand preferences
• Dining frequency and patterns
• Food preferences and dietary restrictions
• Service quality perceptions
• Price sensitivity
• Demographic details

Notable System Fields

• psid: Participant session ID
• ch: Channel identifier
• pspath: Path/survey flow
• rid: Response identifier
• qsamp: Sampling indicator (duplicate of QSamp)

Data Quality Notes

Encoding Information

• All numeric fields are encoded: Use the Datamap to decode values to meaningful categories
• Text fields preserved: Open-ended responses are stored as-is
• Missing values: Fields may contain null values or have specific encoding for "No response"

Timestamp Data

• All timing data is in seconds
• Represents duration from survey start/section start
• Useful for engagement analysis and respondent behavior patterns

Respondent Deduplication

• record field: Primary unique identifier
• uuid field: Secondary identifier for participant tracking
• Use these to ensure unique respondent analysis

How to Use This Skill

1. Basic Data Exploration

Start by understanding the core respondent attributes and geographic distribution:

• Explore the distribution of respondents across countries
• Review the mix of respondents by channel and sampling type
• Check completion status breakdown

2. Demographic Analysis

Use the core fields to segment respondents:

• Filter by country using hCountry field values (1-47)
• Group by channel: Open, Loyalty, Integrated, 3PP
• Segment by sampling type: Rep, Augment, Oversample variants
• Analyze completion status: Qualified, Terminated, Overquota, Partial

3. Survey Response Analysis

Access main survey questions in columns 9 and beyond:

• Use the Datamap to decode numeric responses
• Analyze Likert scale responses for dining preferences
• Cross-tabulate responses with demographics

4. Timing Analysis

Analyze survey engagement using TimeSectionr1-r4 fields:

• Calculate average completion time per section
• Identify sections with highest dropout
• Compare engagement across countries/channels

File Access

Direct Data Access

• Source File: /c/Users/aayus/grey-chain/projects/runner_gpt/2409899.xlsx
• Read Format: Excel .xlsx
• Recommended Tools: Python (openpyxl, polars), Excel, or database imports

Data Export Options

• Convert to CSV for analysis in Python/R
• Import to database for complex queries
• Use Parquet format for efficient analytics

Limitations & Considerations

1. Field Encoding: All numeric fields require Datamap reference for interpretation
2. Column Volume: 14,739 columns makes direct exploration challenging - use field selection
3. Data Size: 7,916 rows × 14,739 columns = large file requiring adequate system resources
4. Regional Variations: Some questions may have localized versions for different countries
5. Version Tracking: Different respondents may have completed different survey versions (dTrack 0-11)

Support & References

Understanding the Data

• Refer to Datamap sheet for complete encoding reference
• All numeric codes are documented with their text equivalents
• Column naming follows pattern: [FieldCode]: Description

Analysis Tips

• Always use the Datamap to decode values
• Filter for "Qualified" status (status = 3) for primary analysis
• Consider survey version (dTrack) when comparing across time
• Use country coding (hCountry) for geographic analysis

Questions or Issues?

This skill provides comprehensive analysis of the 2024 Restaurant Survey dataset. For specific analysis requirements or unusual patterns, refer back to the Datamap sheet for complete field documentation.

---

Skill Generated: October 30, 2025 Data File: 2409899.xlsx Data Version: As of analysis date Last Updated: 2025-10-30