Data-based Prediction Backend

There is a common pattern for consuming and processing raster data that results in a graph or geospatial geometry.

This innovation story aims to achieve 2 things:
1. Build a demo-able data/prediction based tool that is useful (e.g. Tephigram, lightning starts)
2. Given #1 above, have a backend API to be able to quickly spin up other data/prediction based tools in the future.

**Business Values / Outcomes**:
By doing #1, we build towards something that's useful while also discovering what is needed / what is possible for a data/prediction API. By creating #2 we can leverage the work for delivering future tools quicker / have more time to demo/present/get feedback.

**Questions**:
- What SMEs do we need? Do we need someone to describe Tephigrams?
- What data do we need access to? How do we get data for Tephigrams?
- Is tephigram the best starting point, since forecasters already have a tephigram tool.

**Potential Architecture?**

No implementation files, tests, or entry points are named. Start by resolving the proposed demo, required raster and prediction data, data access, and subject-matter expertise, then inspect the existing Python/FastAPI backend for relevant data-processing patterns. Done means a useful demonstration exists and the backend API supports spinning up similar data- or prediction-based tools.

Written by the indexing model from the issue text.