talk on conference website
Grab launched a navigation product for Driver partners in 2018 with the help of OSM. We contributed to OSM data in Singapore and made it navigation ready to enable launching in-app navigation for Grab's driver partners. This talk is a glimpse of the journey to get the map data quality and in turn the product experience to be the best suited one for driver partners in Singapore.
Grab's journey with OSM started in 2015, with the road geometry being the one mostly used to calculate ETAs and routing. Based on this use case, we started enhancing the OSM data using GPS probe based detections and satellite imagery. After the success we saw with the ETA and routing quality that we could achieve with OSM in Singapore, we started dabbling with the idea of piloting Navigation quality maps on OSM.
Our journey on this started with researching the local traffic rules, signboards and restrictions that are relevant for Navigation. As Singapore is one of the countries in SEA which is well marked with all signs in SEA, we then moved on to solve the next level of challenge which is adding street level imagery across the country with the help of different cameras we could lay our hands on for the pilots. At times we used mobile phones with KartaView app and others we managed to add using better quality Go-Pro like cameras. Few interesting learnings along the way; Rain can play a spoilsport to your plans at the last minute; Government, building authorities would need time to give you permits for collecting imagery and you will need to plan those into your collection efforts; Roadworks, Road closures, congestion do not let you cover all streets fully. Despite all the planning, sometimes you may not get permissions, Eg: Private condos. For such situations, you can work with the local community to help add more imagery.
Interesting learnings while mapping Singapore: Tunnels, Underpasses, etc. criss cross on multiple levels. Oftentimes mappers get confused, despite the availability of imagery. Road name abbreviations are sometimes not standardised and will have inconsistency when showing on maps.
Maps are never perfect even in a well mapped country like Singapore as the physical world around you is changing. Instead of trying to be perfect, you can aim to be very agile and responsive, making the real world changes reflect in the map data as close to realtime as you can get. What can help you in doing this well is working closely with local transport authorities (SLA, LTA, etc. in case of SG); Other major realtors and establishments which run several malls, office buildings, tech parks, etc. And last but not the least - community. Both the OSM community and our driver partner community helped us immensely here.
The journey to making the map for Singapore was quite arduous and needed the continued support of OSM community mappers, driver partners, local map operations teams. Everyone constantly keeping the map updated and improved helped us make the navigation to be best suited for the needs of Grab driver partners. We engaged these communities through various activities such as Geo*Stars, mapathons with organisations such as GeoWorks and hosted in-house mapathons to enable understanding and contribute to the amazing OpenStreetMap data.
For our driver partner community, we started off with a whatsapp group to gather feedback and then proceeded to an in-app collection mechanism to make this reporting of road updates, map changes, a lot faster and seamless. This coupled with our on-ground map operations teams which run road tests of the navigation helped us ensure our map stays as updated as possible. Working closely with the community ensured that they also could see the impact they were making on the maps every day. This made Grab Driver partners to like Grab Navigation even more as it empowered them to build it together with us.
This journey doesn’t end here, and we will be working even more closely with the community across South East Asia as this is the best way to scale these efforts on building maps for everyone.