diff --git a/src/etl/extract.py b/src/etl/extract.py index 93997b9..bf32b84 100644 --- a/src/etl/extract.py +++ b/src/etl/extract.py @@ -30,12 +30,4 @@ def save_to_duckdb(df, table_name, db_path): # + if __name__ == "__main__": - # Extract data from URL - url = "https://data.cityofnewyork.us/resource/erm2-nwe9.json" - df = extract_data(url) - - # Save to duckdb - db_path = "data.duckdb" - table_name = "nycitydata" - save_to_duckdb(df, table_name, db_path) - + pass diff --git a/src/notebooks/experimental-work.ipynb b/src/notebooks/experimental-work.ipynb new file mode 100644 index 0000000..5c4108e --- /dev/null +++ b/src/notebooks/experimental-work.ipynb @@ -0,0 +1,21 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# Code to extract data from kaggle" + ] + } + ], + "metadata": { + "language_info": { + "name": "python" + }, + "orig_nbformat": 4 + }, + "nbformat": 4, + "nbformat_minor": 2 +}