Geospatial Impact Analysis of Hurricane Ian on Tweet Sentiment

On Sept. 28, Hurricane Ian made landfall near Cayo Costa in southwestern Florida. In this example, we want to visualize the impact that the Hurricane had in terms of sentiment of Twitter tweets in different distances around the Hurricane path.
We first read in a prepared dataset that contains the timestamp, location, and sentiment of tweets about Hurricane Ian around the world. In this example, we want to analyze only the tweets within a certain distance of the Hurricane path. To generate the different distances from the Hurricane path we use the Multiple Ring Buffer. In addition to the sentiment, we want to compute the affected population per distance ring. To compute the population per distance we combine the population (US2020 Census Data) with the county geometries (US2020 TIGER Map) of Florida (FIPS=12). The population per county geometry is combined with the Hurricane path distances via the Overlay node. Since we are only interested in the distances, we can use the Dissolve node to aggregate the county geometries per distance ring. The dissolved geometries are then used to filter out all tweets that do not intersect with one of the geometries. Finally, the result is visualized on an interactive map.

To make it easier to follow the different geometric operations (buffer, overlay, dissolve) we have added several Geospatial View nodes that depict the different processing steps.

Geospatial Analytics is fully developed in Python, e.g. the Geopandas library, which was heavily used to write the nodes. All the nodes provided with the extension are the perfect toolkit to apply geospatial technologies in a no-code/low-code way, so also beginners can benefit from this kind of analysis. The time travelling visualization is done using the time playback function in the Kepler.gl Geoview node.

External resources

• Request US Census API key (for free)
• Geopandas Documentation
• Center for Geographic Analysis