Results and implications
The SQUID Bike project has demonstrated in this paper, an automated data workflow to interrogate the general conditions of New York City's Bike lane infrastructure. The authors have collected bike lane imagery data in excess of 50 miles consisting of over 5000 distinct images and over 500,000 accelerometer readings.
The results show that with further work, a robust standard to measure bike lane infrastructure for any city in the world can be made possible using a low-cost approach assuming that bike lane imagery is collected in a purposeful manner. Furthermore advances in Computer Vision may also lead to a fully automated inspection process.
The implications of a fully automated inspection process is that cities can collect high quality, ground truth data about their bike lane infrastructure quickly in a cost-effective manner. We argue that adopting such a practice empowers a city to enter into an anticipatory paradigm for bike lane maintenance and allow transportation agencies be more responsive to the 2-wheeled commuter.
Limitations and improvements
Various constraints involving hardware and software are limitations to this work in this current state. Most of which are solvable with further work. Image Quality, Battery life, and accelerometer discrepancies across phone models are some limitations on the data collection side, whereas inaccuracies in snapping points to bike lanes, among other general GPS inaccuracies, are concerns on the data analysis side. In the following paragraph we will explain how these limitations affected our analysis, how we overcame them and, in the rare cases we didn't, how we plan to solve them in the future.
Limitations to data collection
Most of the complexity around collecting data is around the collection of good quality bike lane imagery. While using a purposeful mount helped in acquiring stable imagery, going over very rough stretches of road or going over deep potholes created vibrations enough to loosen the mount and introduce additional movement of the phone which causes distortions to the imagery and accelerometer data. These situations were few and far between and constituted a small number of data points. Upon using a stable moun, this issue was successfully mitigated.
Recording images using OpenStreetCam consumes a lot of battery charge. A 30 min bike trip tends to drain close to half the battery of a standard Android phone. For this reason, the crowdsourcing approach for collecting data may be limited. Using an external battery pack will mitigate this isue but will require additional mounting on the bicycle.
Correcting for variance in accelerometer readings for the same defect from different devices is another limitation. However, these inconsistencies weren't high and the overall magnitudes for the accelerometer readings tended to be similar. This maybe mitigated by assuming some calibration settings for commonly used phones.
Limitations to analytics
Limitations to analytics include the precision of GPS data and the quality of the underlying shapefile. The naivete of the computer vision remains as the single largest limitation which can be improved with further maturation of computer vision and image processing techniques.
Conclusion
Expanding and maintaining a safe and comfortable bike lane infrastructure is key to supporting a progressive and climate friendly transit agenda. We believe that cities need to quickly adapt their core operations to deliver sustainable transit futures. Maintaining a sizable network of bike lane infrastructure is key to achieving these goals and contributing to overall improvements in quality of life indicators. To that end, we believe that SQUID-Bike is very relevant to realizing transit goals by empowering city agencies be more responsive and do more with fewer resources. Furthermore SQUID-Bike is also a tremendous opportunity for cities to engage with multiple stakeholders and pioneer a participatory inspection process where everyday citizens can be directly involved and actively contribute to city maintenance operations.
Author contributions
Felipe Gonzalez worked on data engineering (data gathering, cleaning, snapping function, overall coordination of python's processes)
Nicola Macchitella worked on data engineering (data gathering, GIS analysis and overall coordination of python's processes)
Geoff Perrin worked on the computer vision code (lane markings and defects classification), Tableau visualization, as well as setting up the amazon web service database and the code that pushed our data to the database.
Sichen Tang worked on data analysis to profile NYC's Bike Lane infrastructure, implement computer vision techniques (color score) in addition to implementing other front end components.