-
Notifications
You must be signed in to change notification settings - Fork 0
Home
The climate and health interactive platform allows users to explore the data generated in a machine-learning assisted rapid map of the evidence on climate and health, conducted by the University of Leeds, the London School of Hygiene and Tropical Medicine, and the Mercator Research on Global Commons and Climate Change in Berlin. The evidence map was produced for the UK Department for International Development.
The map coded 15,000 scientific papers by subject (using topic modelling) and by place of study (using a geoparser). The interactive platform allows users to explore this data.
https://apsis.mcc-berlin.net/climate-health/
The regions section displays 6 geographical regions with countries that are of special interest to DfID. In each country, we place a dot in the location of all studies for which we found a location in that country. The bar chart to the right of the map describes the topics which are highly prevalent (compared to the global average) in the region. The sidebar on the right shows the documents and topics that are currently selected. By default this will show all documents in the countries of interest to DfID.
The following actions can be taken to select or view different data
- Use the blue slider to scroll down and view less common and underrepresented topics in the region.
- Select (by left-clicking and dragging on the map) a group of locations on the map to view the documents mentioning these locations in the table on the right. Doing this will also update the bar chart to show the distribution of topics in the selection made.
- Click on a topic bar to view the documents in the region with the highest scores for the selected topic.
This "heatmap" shows the number of documents in various combinations of topic category. The columns refer to groups of topics which describe climate drivers, while the rows refer to groups of topics which describe health impacts. Each cell contains the number of documents which have high topic scores for the climate driver in the column and the health impact in the row.
Darker colours refer to higher numbers, but we may also want to highlight large and small values in the context of each row or each column. Clicking on "Normalise by column sum" adjusts the colour scale so that the colours reflect each cell's proportion of the column sum, and "Normalise by row sum" does the same for the row sum.
Clicking on a cell updates the table on the right side of the screen to show the documents which refer to topics in the cells column and row. For example, the screenshot below shows the documents relevant to Infections disease & Temperature
The topic map represents each document with a dot placed according the document's combination of topics. The two axes are arbitrary, but the 2-dimensional representation aims to put similar documents close together.
Hover over a dot to see the document title appear at the bottom-left corner of the screen. Double-clicking on a dot will search for the document online. The tabs on the right-hand side of the screen give more options for displaying and filtering the data
- World View shows the same documents on a geographical map of the world. Hovering and clicking works in the same way
- Draw labels draws labels for each topic that point to clusters of documents related to the topic
- Continent Filter lets you filter the data by continent