Name
Abstract | ||
|---|---|---|
As cities undergo rapid sprawl and urbanization, it is commonly becoming a difficult task to find public toilets available. Building new public toilets becomes a promising way to alleviate this issue. However, where to build new public toilets is challenging. Traditionally, urban planners rely on empirical experience and surveys to understand the local toilet demand, which is unreliable and also time and labor-consuming. In this paper, we propose a data-driven approach to tackle the site selection problem of public toilets. Specifically, we propose a two-phase framework to discover knowledge from existing public toilets and use them to guide future planning and construction. In the first phase, we identify the candidate areas for new public toilets based on human mobility, land use, urban structure, and etc. In the second phase, we propose a learning-to-rank method to predict the demand level of the candidate areas and identify the optimal sites for placing new public toilets by simply ranking. Our approach combines the geographic characteristics of the city with mobility patterns of human activity by considering human activity, area functionality, road network, and the toileting demand of taxi drivers. Finally, we evaluate our approach by using multiple datasets including the taxi GPS trajectory data, POI data, and road network data in the real world from the city of Chongqing, China. The experimental results demonstrate the effectiveness of our proposed method. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1109/SmartWorld-UIC-ATC-SCALCOM-IOP-SCI.2019.00217 | SmartWorld/SCALCOM/UIC/ATC/CBDCom/IOP/SCI |
DocType | Citations | PageRank |
Conference | 1 | 0.38 |
References | Authors | |
0 | 6 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Chaoxiong Chen | 1 | 1 | 1.73 |
| Yuyang Liu | 2 | 1 | 0.38 |
| Chengwu Liao | 3 | 1 | 1.06 |
| Chao Chen | 4 | 1 | 2.07 |
| Liang Feng | 5 | 601 | 48.54 |
| Zhu Wang | 6 | 23 | 8.48 |