`handbag_data/us_government_bigquery/BIGQUERY_SETUP_GUIDE.json`

69 lines
{
  "title": "Google BigQuery Public Datasets for Handbag Data",
  "description": "Access fashion/product datasets via BigQuery",
  "datasets": [
    {
      "name": "Google Open Images",
      "description": "9M+ images with annotations",
      "handbag_relevance": "Product images, fashion items",
      "access": "bigquery-public-data.open_images",
      "query_example": "SELECT * FROM `bigquery-public-data.open_images.annotations` WHERE label_name LIKE \"%bag%\" OR label_name LIKE \"%handbag%\" OR label_name LIKE \"%purse%\""
    },
    {
      "name": "Fashion MNIST",
      "description": "70,000 fashion product images",
      "handbag_relevance": "Fashion classification dataset",
      "access": "Via Kaggle/TensorFlow",
      "notes": "Already downloaded in many datasets"
    },
    {
      "name": "Common Crawl",
      "description": "Web crawl data",
      "handbag_relevance": "E-commerce product data, fashion sites",
      "access": "bigquery-public-data.common_crawl",
      "notes": "Requires processing for handbag-specific data"
    }
  ],
  "setup_steps": [
    "1. Create Google Cloud Project:",
    "   https://console.cloud.google.com/",
    "",
    "2. Enable BigQuery API",
    "",
    "3. Install Google Cloud SDK:",
    "   curl https://sdk.cloud.google.com | bash",
    "",
    "4. Authenticate:",
    "   gcloud auth login",
    "   gcloud config set project YOUR_PROJECT_ID",
    "",
    "5. Query public datasets (free up to 1TB/month):",
    "   bq query --use_legacy_sql=false '",
    "   SELECT * FROM `bigquery-public-data.open_images.annotations`",
    "   WHERE label_name LIKE \"%bag%\" LIMIT 1000",
    "   '"
  ],
  "handbag_relevant_queries": [
    {
      "dataset": "Open Images",
      "query": "SELECT\n  image_id,\n  label_name,\n  confidence,\n  xmin, ymin, xmax, ymax\nFROM `bigquery-public-data.open_images.annotations`\nWHERE\n  (label_name LIKE '%bag%'\n   OR label_name LIKE '%handbag%'\n   OR label_name LIKE '%purse%'\n   OR label_name LIKE '%clutch%')\n  AND confidence > 0.8\nLIMIT 10000"
    }
  ],
  "cost_estimate": {
    "query_cost": "Free for first 1TB/month",
    "storage_cost": "Results stored in your project (minimal)",
    "download_cost": "Egress charges may apply for large downloads"
  },
  "alternative_access": [
    "Kaggle Datasets (pre-downloaded): https://www.kaggle.com/datasets",
    "TensorFlow Datasets: https://www.tensorflow.org/datasets",
    "Hugging Face Datasets: https://huggingface.co/datasets"
  ],
  "notes": [
    "BigQuery best for large-scale image datasets",
    "Open Images has 9M+ product/fashion images",
    "Requires Google Cloud setup (15-30 minutes)",
    "Free tier sufficient for this project",
    "Consider Kaggle/TF Datasets for easier access"
  ]
}