Using NASA Data Instead of Commercial Satellite Imagery

January 31, 2025

We are currently exploring data-driven business opportunities using publicly available Earth observation data.
Instead of relying on commercial satellite imagery, this study focuses on open NASA datasets and an experimental data-processing workflow generated with the assistance of GPT-4.

This article documents an early-stage prototype and is intended to share technical observations rather than present a finished commercial product.

Objective

The objective of this experiment was to evaluate whether publicly available NASA data could be processed and visualized in a way that supports future analytical services.

One potential use case under consideration is large-scale infrastructure planning, such as identifying offshore areas of interest and correlating them with on-land networks.
At this stage, however, the goal is limited to technical feasibility testing.

Approach

Access to NASA’s public data portal was obtained through user registration.
Large satellite image files (GeoTIFF format) were downloaded locally due to bandwidth and server-load constraints.
Python scripts were generated with GPT-4 assistance to load, resize, and visualize the data.
The focus was on handling very large image files safely and reproducibly.

Example: Image Processing Workflow (Python)

Example: Image Processing Workflow (Python)
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image, ImageFile

Disable pixel limits for large satellite images

Image.MAX_IMAGE_PIXELS = None
ImageFile.LOAD_TRUNCATED_IMAGES = True

LOCAL_TIF_FILE = “MOD44WA1.A2023001.061.Global-WGS84.500.2024074204203.tif”

def load_local_tif():
try:
with Image.open(LOCAL_TIF_FILE) as img:
scale_factor = 0.1
new_size = (
int(img.width * scale_factor),
int(img.height * scale_factor)
)
img_resized = img.resize(new_size, Image.LANCZOS)
return np.array(img_resized)
except Exception as e:
print(f”Failed to load TIF file: {e}”)
return np.zeros((100, 100, 3), dtype=np.uint8)

This script loads a locally stored GeoTIFF file, reduces its resolution to 10%, and prepares it for visualization.
The purpose is to validate data handling, not to optimize image quality at this stage.

Observations

The spatial resolution of the selected public dataset is insufficient for detailed commercial analysis in its raw form.
Visualization is technically possible, but further preprocessing and geospatial refinement would be required.
Public datasets are suitable for concept validation and prototyping, but careful dataset selection is critical.

Limitations

This experiment does not evaluate proprietary or high-resolution commercial satellite data.
No automated API-based data ingestion is implemented yet.
Analytical value extraction (beyond visualization) is outside the current scope.

Next Steps

Evaluate alternative NASA datasets with different resolution characteristics.
Introduce geospatial alignment and metadata-aware processing.
Transition from visualization to quantitative analysis.
Continue documenting findings in English for an international audience.

Conclusion

This work represents an early technical exploration of using NASA’s open data for future data-driven services.
While the current results are not commercially deployable, they provide valuable insight into feasibility, constraints, and required next steps.

Further development will focus on data quality, processing accuracy, and practical applicability.