NYCMotorVehicleAccidents

The GET Request retrieves the current number of records from the API,

The JSON Path node extracts this value from the JSON response.

The Math Formula calculates the number of required API requests by dividing the total row count by 50,000 and rounding up.

Finally, Table Row to Variable converts the calculated number of pages into a flow variable that controls the subsequent loop.

The Counting Loop Start uses the calculated number of pages to determine how many times the API request must be repeated.

During each iteration, Math Formula (Variable) calculates the offset by multiplying the current iteration number by 50,000. 

String Manipulation (Variable) then inserts this offset into the API URL, ensuring that each request retrieves the next batch of records.

The GET Request retrieves one batch of up to 50,000 collision records using the dynamically generated API URL. 

JSON Path, Ungroup, and JSON to Table extract the individual collision objects, create one row per collision, and convert their attributes into table columns.

Finally, the Column Filter removes unnecessary API metadata before the processed batch is passed to the Loop End.

The imported string values are converted into appropriate date, time, and numerical data types to prepare the dataset for analysis.

A review of the final data-cleaning node confirmed that no duplicate collision_id values were present.

No further cleaning was performed at this stage, as additional transformations depend on the specific analysis and its requirements.