A11y: Use AI tools to describe "ItemViewer" maps/images/data layers

[ewlarson]

For example, given a WMS item layer preview:

We could ask ChatGPT to describe a screenshot of the preview alongside the item's title: "2005 Rural Poverty GIS Database: Uganda" and description:

The image you provided appears to be from a web-based geographic information system (GIS) interface, titled "2005 Rural Poverty GIS Database: Uganda." Here are some key details based on the map content and the description title:

1. **Geographical Coverage**: 
   - The map highlights Uganda extensively, with the detailed blue line grid covering most of the country, representing rural areas.
   - Surrounding regions of Kenya, South Sudan, and parts of the Democratic Republic of Congo are visible but not highlighted with the same detail, suggesting the focus is specifically on Uganda.

2. **Map Features**:
   - **Zoom and Layer Controls**: There is a vertical zoom control on the left side of the interface, allowing users to adjust the map view for closer inspection of specific areas. A transparency slider is also present, set at 75%, possibly controlling the visibility of the overlaying data on the base map.
   - **Base Map and Overlay**: The base map is minimal, primarily showing country boundaries and major city markers. The overlay with detailed blue grid lines likely represents administrative or statistical boundaries related to rural poverty data in Uganda.
   - **Interface Tools**: The map interface includes zoom buttons and a settings or layers button (the square grid icon), indicating customizable view options.

3. **Purpose and Data Visualization**:
   - Given the title, "2005 Rural Poverty GIS Database: Uganda," the map's primary purpose is to visualize rural poverty data from 2005. The detailed grids could represent various poverty metrics, such as population statistics, income levels, access to resources, or other relevant socioeconomic data.
   - The use of an overlay suggests the data could be layered to compare different information sets, like health, education, or infrastructure, against the backdrop of rural poverty.

4. **Interactivity and Functionality**:
   - The presence of interactive controls suggests that the map allows users to explore different data layers or adjust the map for a more detailed examination of specific regions or data points.

5. **Data Source and Attribution**:
   - The map credits "Leaflet | OpenStreetMap contributors, © Carto" at the bottom right. This indicates the use of Leaflet for map interaction, data from OpenStreetMap contributors for the geographical details, and Carto for the map's styling and possibly the data visualization tools.

This map serves as a vital tool for analyzing and understanding the distribution and dynamics of rural poverty in Uganda as of 2005, allowing for in-depth analysis and potentially aiding in policy making and resource allocation.

We could then include this alongside the Leaflet/OpenLayers map as the accessible text version.

[karenmajewicz]

Great idea that I had not considered. I wonder how common this is? I did a search for "ai for image alt text" and found several online apps.