The talk discusses the challenge of incorporating real-time traffic speed information into open-source routing systems that use OpenStreetMap (OSM). Most open-source routing engines rely on static driving speeds for different road types due to the lack of comprehensive and open traffic speed data. The presentation introduces a method to model hourly traffic speeds for street networks in ten global cities and their integration into the openrouteservice routing engine. Existing traffic speed datasets often lack openness, consistent formats, or OSM integration. Uber Movement offers a valuable dataset, but it only covers roads with sufficient Uber user data. The study proposes a model for traffic speed based on OSM tags, an adapted betweenness centrality indicator, and Twitter data. A gradient boosting regression model is trained and evaluated using Uber traffic speed data as reference. The model's performance is assessed using metrics like R2, RMSE, and MAE. To use the modelled traffic speed data, an experimental traffic integration is implemented in openrouteservice. The effect of external traffic speed data on travel time estimation is evaluated by comparing it to the Google Routing API and the original openrouteservice. The study emphasizes the need for further research on transferability, deep learning approaches, and integrating data from other social media platforms as Twitter has become a paid service. The presentation highlights the potential of leveraging open data and open-source tools for addressing real-time traffic challenges in routing services.