fix(bakery): fix demographics table column shift and define dataset description

[n-issei-777]

Description

This PR fixes a data alignment bug and a missing environment/script variable definition in the launchmybakery BigQuery setup script.

1. Fix Demographics Table Column Shift (Data Corruption)

• The Issue:
  The source CSV file data/demographics.csv contains 8 columns in the following order:
  zip_code,city,neighborhood,median_household_income,total_population,median_age,bachelors_degree_pct,foot_traffic_index
  However, the CREATE TABLE schema in setup_bigquery.sh only defined 7 columns, completely omitting the 4th column median_household_income.
  Because bq load performs a positional mapping for CSV imports, and --ignore_unknown_values=true was specified, the data was shifted left sequentially:

• median_household_income (4th CSV column) was loaded into total_population (4th schema column).

• total_population was loaded into median_age (causing values like 33626 to be stored as the age).

• median_age was loaded into bachelors_degree_pct.

• bachelors_degree_pct was loaded into foot_traffic_index.

• The actual foot_traffic_index (8th CSV column) was entirely ignored and lost.

• The Fix:
  Added median_household_income INT64 in the 4th position of the CREATE TABLE query inside setup_bigquery.sh to perfectly align with the CSV columns and restore data integrity.

2. Define Dataset Description Variable

• The Issue:
  The bq mk command uses --description "$DATASET_DESCRIPTION", but the variable $DATASET_DESCRIPTION was never initialized anywhere in the script, resulting in an empty dataset description.
• The Fix:
  Defined DATASET_DESCRIPTION="Dataset for MCP Bakery Demo" at the top of the script.

[n-issei-777]

Resolved by #53, so this PR is no longer needed.
Closing this PR.