Background Geographically accurate travel routes are necessary to estimate exposure to the environment and its potential influence on travel behaviour. Although assessing travel behaviours with Global Positioning System (GPS) receivers is increasingly common, these protocols place noticeable burden on participants and processing these data is time consuming and error-prone. Interactive online mapping surveys allow users to draw their own travel routes, and may offer a time and cost-effective alternative; however, these routes are still self-reported, and their true accuracy remains unknown. Methods A total of 196 adolescents drew their usual route to school within an online mapping survey and wore a GPS receiver for 7 days. Individual home-to-school routes were extracted from GPS data. Generalized linear mixed models were used to assess differences in distance and spatial agreement between routes, and how these varied by mode of travel and other route characteristics. Results GPS-assessed routes were longer than the routes participants drew across all travel modes. Routes travelled actively displayed 12.32% higher agreement compared to those travelled passively (p < 0.01). Taking multiple routes to school (29.9% of participants) reduced the agreement by 10.76% (p < 0.01). Every additional travel mode transition (e.g., during multimodal trips) was associated with 2.20% lower agreement (p < 0.01). In total, 40.7% of participants used more than one travel mode to school over the assessment period. Conclusions Online mapping surveys are a feasible method for route assessment in adolescents, particularly for active travel routes. With the integration of survey questions, there is considerable potential for understanding the intricacies of travel behaviours. However, the self-reporting error seems more pronounced for longer routes, and when multiple travel modes are used. Researchers should consider the advantages (e.g., ease of collection) and disadvantages (e.g., lack of temporal information) when deciding if the data obtainable are sufficient to answer their research questions.