Data: Land Transport

As a first step in contributing to improving transport planning in cities (Philippines as a priority), you may find a collection of historical transport network datasets in this website. I had extracted 2015 and 2024 datasets from OSM through the ohsome historical database facility. The files for each city comprises of static images, geojson files which you can directly use in GIS programs such as QGIS, as well as summaries in excel (lengths) for various types of infrastructure (primarily different types of roads).

Of course there are limitations, as these are basedon OpenStreet Map (OSM) data. At its best, OSM offers a freely accessible, continuously updated, and impressively detailed global map database—a democratic counterweight to proprietary geospatial platforms. Its granularity, particularly in urban areas and developing regions with active contributor communities, allows planners, researchers, and NGOs to trace roads, footpaths, bus stops, and bike lanes that might not appear in official records. For cash-strapped institutions, it offers a viable alternative to expensive commercial datasets.

But OSM’s open model is both its strength and its Achilles’ heel. Coverage is uneven, reflecting where contributors are active rather than where needs are greatest. Rural areas, lower-income countries, and less digitally connected communities are often underrepresented. Data quality varies: while some areas boast high precision and tagging consistency, others suffer from outdated entries, inconsistent schemas, or gaps in infrastructure types. Importantly, because OSM is built for general-purpose mapping rather than specialized infrastructure inventories, it may lack the technical rigor or classifications demanded by engineers or national transport agencies.

It is a powerful starting point—flexible, scalable, and transparent—but not a panacea. It serves best when used critically, supplemented with local knowledge or official records, and understood for what it is: a civic commons, not a certified dataset.