Open each app in your browser to explore features and data. If you notice any issues please drop me a line!
1. Cluster Explorer
https://simonabisiani-clustering-analysis-dashboard.share.connect.posit.cloud/
2. Outlet Map
https://0198a3be-1c47-dc00-7fcf-70907dd8e320.share.connect.posit.cloud/
Purpose:
- Explore distribution of location mentions across articles
- Cluster (by publisher or otherwise?)
- See if location occurrence makes articles “clusterable”
- Could layer other article-level properties
Elements:
<aside>
- Menu → top left (1/3)
- Outlet selector
- Cluster selector
- Metric selector
</aside>
<aside>
- Stats / Plots → bottom left (2/3)
Default:
- Textbox: Corpus Info, including umber of outlets, number of publishers, number of locations and articles.
- Table:
- Cluster → Rural (n = 5), Hyperlocal (n = 10)
- GINI → mean, median, sd
- Convex, Radius → mean, median, sd
- LSOA count → mean, median, sd
- Moran’s I → mean, median, sd
- Plots:
- Boxplots / ggridges
- Heatmap
- Cumulative distribution curves
If outlet is chosen:
- Textbox: Outlet info, including Name, URL, owner, description.
- Table:
- Cluster → Rural
- GINI → 0.2 (rank: 20/360)
- Convex, Radius → 20km2, 30km, (rank: 20/360)
- Top Location → Birmingham (n = 45)
- LSOA count → 16, (rank: 20/360)
- Moran’s I → 1.5, (rank: 20/360)
- Location count → 73, (rank: 20/360)
- Plots:
- Boxplots / ggridges
- Heatmap
- Cumulative distribution curves
</aside>
<aside>
- Map → right
- Locations as scatter on map:
- Size by frequency
- Colour by cluster
- Draw convex hull
- Draw radius around primary location
- Overlay LSOA:
- Colour by frequency
- Colour by type
- Draw radius of 10% locations, and so on, from outlet central location
</aside>
Compare Outlet Section
Spatial Heterogeneity