Findings
Gap Analysis of Roads and Settlementfrom OSM and HDX
- HOTOSM datasets generally have more features and kilometers of roads than OCHA datasets. -- However OCHA datasets equal HOTOSM in "main roads" and "minor roads" categories. It is only in "unclassified" and "paths"that HOTOSM has more features than OCHA.
- HOTOSM datasets have a higher percentage of features with surface information, but a lower percentage of features with names.
- Somalia's HOTOSM dataset has the highest number of features (298,873) and kilometers of roads, while Ethiopia's OCHA/WFP dataset has the oldest data (from 2016).
- Ethiopia's HOTOSM dataset has the most languages (6). Most of the data is in English.
- There is a lack of consistency in the name columns accross UN datasets.
- Many datasets are only available in shapefile format, with limited availability in other formats.
- HOTOSM datasets are more comprehensive in terms of format availability, with over 10 optiosn available including sometimes tiles, Garmin, etc.
- HOTOSM datasets generally have more features and attributes than OCHA datasets, but OCHA datasets are more consistent.
- HOTOSM datasets have a higher percentage of features with names, but OCHA datasets lanaguage coverage is more consistent.
- Ethiopia's HOTOSM dataset has the most languages (6), but most of the data is in English.
- The distribution of settlements is very different between HOTOSM and OCHA datasets, indicating room for improvement in both datasets.
- The count of unique features is deceiving, suggesting that there may be inconsistencies in the data.
- HOTOSM datasets appear to be using different borders, including areas in the center of Sudan that are not included in OCHA datasets.
- Somalia's HOTOSM dataset has the highest number of roads features and kilometers of roads -- which indicates that there is a need to improve the quality of the data.
- Ethiopia's HOTOSM dataset has the most languages, but most of the data is in English.
- Ethiopia's OCHA/WFP dataset has the oldest data (from 2016) -- OSM data is always more dynamic.
- HOTOSM datasets tend to have more features and kilometers of roads, but OCHA datasets tend to have more attributes and be more consistent.
- OCHA datasets tend to have a higher percentage of features with names, especially in the local language.
- HOTOSM datasets tend to have more languages, but most of the data is in English.
- The two datasets have different strengths and weaknesses, and could potentially be used together to create a more comprehensive dataset.
- Inconsistency in the URL names on HDX for UN exports.
- Settlements and populated places might have different definitiosn (?)