Sudan Settlements

HOT Populated Places vs HDX COD Settlements

The Area and The Data

Findings:

Populated Places / Settlements

• HOT features seem to have a more compact distribution and better correspondence with human structures, but OCHA's coverage is broader.
• OCHA's data has more attributes (31 vs 9) and 100% of features have no empty cells.
• HOT's data has a lower percentage of features with names (37.51% vs 100% for OCHA, but OCHA's names are in a single column).

⠀Key Attributes:

• HOT's key attributes include name (10,284), name:ar (5,709), and source (7,154)
• OCHA's key attributes include name, p-codes (admin1-2), and old/new admin values.

⠀Improvement Plan:

• Process name columns to make "Name:" exclusively for local language and "name:en" for English.
• Potentially do transliterations.
• Validate and potentially import features from OCHA using population estimates, building layers, or other proxies.
• Determine if there is any meaningful difference between settlements and populated places as concepts that explain the differences

Roads:

• HOT has less than 1% of roads with a name and 25% with surface info, but could have up to 300,000 km of roads that OCHA doesn't have.
  • HOT roads are mostly unpaved and rural
• OCHA has fewer features and only one column, but could have 17,000 km of roads that HOTOSM doesn't have.

⠀Key Attributes for Roads:

• HOT's key attributes include name (2,769), name:ar (1,735), and source (38,877).
• OCHA's key attributes include name.

⠀Improvement Plan for Roads:

• Process name columns to make "Name:" exclusively for local language and "name:en" for English.
• Potentially do transliterations.
• Validate and potentially import features from OCHA using meta data and rapid if necessary. Put special interest in main roads