{
    "table_name": "startup_funding",
    "table_type_tags": [
        {
            "tag": "event",
            "confidence": 0.6
        },
        {
            "tag": "dimension",
            "confidence": 0.3
        },
        {
            "tag": "path",
            "confidence": 0.22
        }
    ],
    "primary_type": "event",
    "insufficient_evidence": false,
    "entity": {
        "columns": [
            "Startup Name"
        ],
        "unique_per_row": false
    },
    "time_axis": {
        "columns": [
            "Date dd/mm/yyyy"
        ],
        "kind": "event_time"
    },
    "reasoning": "Each row is a funding deal that happened to a startup on a given date. Startup Name repeats (RPUV 1.24, Uniqueness 0.808, Ola Cabs and Swiggy with 8 rows each) and Date dd/mm/yyyy is a full event timestamp (1,035 distinct values, stored as text). Sr No (Uniqueness 1.000) is only a row counter, not a catalog ID. Replaying the rows would not reconstruct a clean entity state: Amount in USD is 'N/A' in 959 rows and is messy text, so the rows read as bare occurrences, which points to event. The dimension and path readings are weaker alternatives. Dimension applies because Sr No is unique per row and each record carries descriptive attributes. Path applies because funding stage and cumulative funding could loosely be replayed."
}