synthetic_customer_demand_prompt.txt

You have a CSV file called `locations.csv` with columns: name, longitude, latitude, type (including 'Customer' rows), DCs, and plants. 

I want you to:

1. Filter the data to only include rows where `type == 'Customer'`.

2. Generate synthetic one-period demand for these customers:

   - Normal scenario: Draw from a normal distribution (mean=100, std=20), clip negatives at 0.

   - Lognormal scenario: Draw from a lognormal distribution (mean=4.5, sigma=0.5).

3. Compute a local density factor:

   - Convert lat/lon to radians.

   - Use a BallTree with haversine distance to count the number of other customers within 100km.

   - Compute density factor = (neighbors+1)/(average_neighbors+1).

4. Apply this density factor to both normal and lognormal demands, producing four scenarios:

   - Normal Base

   - Normal Density-Scaled

   - Lognormal Base

   - Lognormal Density-Scaled

5. Create four separate Folium maps (one per scenario), each saved as an HTML file. 

   - Markers represent customer locations.

   - Marker size is smaller, scaled by demand (smaller than before).

   - Initial map zoom is 4.

   - Color markers based on demand volume (Low: <80 green, Medium: 80-120 orange, High: >=120 red).

6. Arrange all four Folium maps side-by-side in a 2x2 grid in a single Jupyter cell using iframes and a simple HTML block (no ipywidgets).

7. Finally, create histograms of each demand scenario (Normal Base, Normal Density-Scaled, Lognormal Base, Lognormal Density-Scaled).

Please provide a single Python code cell that executes all the above steps and displays both the maps and the histograms in a Jupyter notebook environment.