Sajjad Anwar

The geospatial world is moving to cloud-first and cloud-native approaches. Movements like STAC and COG have transformed how people use raster data in the last couple of years. OpenStreetMap has much to gain from thinking about different cloud infrastructure architectures. This talk will discuss what it takes to run the OpenStreetMap ecosystem in the cloud and present the history and work on a project called [OSM Seed](https://github.com/developmentseed/osm-seed/). We learned so much while building OSM Seed and think it can be a blueprint for running OpenStreetMap on cloud infrastructure.

The OpenStreetMap ecosystem contains many open source software projects beyond the Rails application. This includes tools that are used by mappers on a daily basis, like, Overpass, Tasking Manager, iD, JOSM, tile servers, data processing applications, Nominatim and so on. It’s a complex ecosystem of growing tools. These are maintained by different individuals and organisations. OSM Seed started in 2018 after several attempts to containerise OSM software for easy installation. [I talked about this vision](https://2015.stateofthemap.us/openstreetmap-software-for-more-than-openstreetmap) in 2015 at State of the Map US. Since then, OSM Seed has grown to be a mature and heavily tested software project that now powers projects like OpenHistoricalMap and data projects at Humanitarian OpenStreetMap Team. We learned so much while building OSM Seed and think it can be a blueprint for running OpenStreetMap on cloud infrastructure. OSM Seed uses cloud agnostic approaches through Kubernetes to bring together projects that make up what OSM is today. These projects continue to get updates and are maintained how they always have been. OSM Seed provides an interface to bring the ecosystem a little closer through better infrastructure orchestration. In this talk, I’d like to discuss a brief history of OSM Seed, some of our lessons trying to move OSM software to the cloud, and open a conversation on what could be useful for OSM going forward.

Over the last few months, we’ve been building an offline first field mapping tool for the OpenStreetMap ecosystem called Observe. Observe makes field surveying, and verification easy for mappers, and works on iOS and Android. My talk will cover how Observe integrates with mapping workflows and helps improve OSM data from the field. I'll also touch upon fundamentals of building Observe and our assumptions on field mapping.

Field verification is an important part of keeping OSM data accurate. So far, field mapping exercises are largely manual, cumbersome, or requires internet connectivity. OpenStreetMap has an active mobile editing ecosystem, but they don't offer the same editing experience as iD for beginners. Most often mapping campaigns need an equally good tool that allows edits from the field to verify existing data and improve data quality. Observe is a cross-platform, offline-first field mapping tool for OpenStreetMap, perhaps the first of its kind. Our primary goal was to build an application that makes field observation easy, and provide comparable experience to iD on Android and iOS — with some success. Observe focuses on browsing OSM data, and allows users to add new points, or verify existing information. The edits made offline are stored on the phone, and uploaded when the mapper goes online. Observe is a product of several iterations of user research and a couple of years of conceptualization. The talk will: - introduce motivations behind building Observe - look at some specific use cases in campaign management - share internals of building the application - discuss challenges around offline editing and our approach to conflict resolution - gather feedback and launch a developer edition of the application