This paper has two aims: to explore approaches to the measurement of children's daily travel to school in a context of limited geospatial data availability and to provide data regarding school choice and distance travelled to school in Soweto-Johannesburg, South Africa. The paper makes use of data from the Birth to Twenty cohort study (n?=?1428) to explore three different approaches to estimating school choice and travel to school. First, straight-line distance between home and school is calculated. Second, census geography is used to determine whether a child's home and school fall in the same area. Third, distance data are used to determine whether a child attends the nearest school. Each of these approaches highlights a different aspect of mobility, and all provide valuable data. Overall, primary-school-aged children in Soweto-Johannesburg are shown to be travelling substantial distances to school on a daily basis. Over a third travel more than 3 km one way to school, 60% attend schools outside of the suburb in which they live, and only 18% attend their nearest school. These data provide evidence for high levels of school choice in Johannesburg-Soweto, and that families and children are making substantial investments in pursuit of high-quality educational opportunities. Additionally, these data suggest that two patterns of school choice are evident: one pattern involving travel of substantial distances and requiring a higher level of financial investment and a second pattern involving choice between more local schools, requiring less travel and a more limited financial investment.