The original goal was to apply machine learning to the accumulated precipitation in the upstream of a river and the water level in the downstream of a river from satellites and to be able to estimate the water level in the downstream from the precipitation. I consulted with a mentor and found that there was already something that would do the job without me having to create it myself: JAXA's Today's Earth.
The goal was reset to visualize facilities that would be affected by floods by synthesizing Today's Earth's predictions of river flooding areas on top of the map service.
Today's Earth provides data in NetCDF format, which is binary oriented and not suitable for web client side development. I tried to implement it in a native application, but time was running out, so I had to make a presentation only.
Since all of the mentors in this study used satellite data in their research, I chose the topic to get their help. By looking at the actual satellite data, I selected data that would be useful in achieving our goals. I would like to examine more examples to confirm the usefulness of the data.
JAXA Today's Earth
JAXA Today's Earth
NASA Panoply