Skip to content

luiscape/unosat-product-scraper-analysis

Repository files navigation

UNOSAT Geo-preview Assessment

Series of scripts that analyze UNOSAT datasets on HDX.

Build Status Coverage Status

Results and Recommendations

Figure 1: shapefiles per resource

Figure 1 above shows that at least 16 resources contain single shapefiles (12% of UNOSAT datasets). That means that, on their respective datasets, a geo-preview will very likely not cause confusion to the users. For a more detailed review, please refer to the analysis page.

As a result, I see two possible actions:

  1. Manually edit datasets: activate the geo-preview only on those datasets in which the resources have a single shapefile. This would include editing ~16 datasets and keeping watch for future additions.

  2. Expose Geo-preview Through API call: at the moment the geo-preview is automatic. It detects resources that could be previewed and activates the geo-preview without user intervention. If we added a property to a dataset (i.e. geo_preview: false), an user could have more control over what to preview and what not to preview. This could be coded into the scraper and configured to only activate the geo-preview on those datasets that have a single shapefile per resource.

1 is short-term; 2 is medium / long-term.

2 is preferred as I believe this kind of issue could become more and more common.

Installation and Usage

Use the Makefile:

run:
  bash bin/run.sh;

test:
  bash bin/test.sh;

setup:
  bash bin/setup.sh;

Those will produce results inside the output/ folder.

Editing Analysis File

The analysis output is generated with a combination of R scripts and R markdown. Assuming you have R installed, run the following:

$ make install_edit
$ make edit

A live editor will start, allowing you to to edit the analysis report.

About

Scripts to analyze UNOSAT products.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published