Coord v. ArcGIS - Curb Data Collection and Regulation Translation
We’ve built Coord to make curb inventorying, optimization and management as seamless as possible, all while maintaining data quality. We’ve run head to head tests in the past to determine how our tools stacked up to other potential solutions being used by cities and engineering firms, but we’ve added a number of features in the last 6 months and wanted to refresh this. The following are the results from this head to head comparison.
For the study area, we selected a complete New York City block, right near our offices by Grand Central Terminal. We’re in Midtown, so there are not as many features or signs that need to be interpreted as you might encounter in a downtown or major nightlife area. In those areas, which typically have more curb features (signage, curb paint etc.), you’d expect some of the steps to take even longer in ArcGIS, especially the collection portion.
We compared three separate collection tools in this test: the Coord Collector app, Collector for ArcGIS without a measuring wheel and Collector for ArcGIS with a measuring wheel.
For the collection piece of the process, the Coord Collector app was about 2.5x as fast as Collector for ArcGIS, although with a major caveat: because Collector for ArcGIS relies on GPS, the quality of the ArcGIS data in an urban environment comes with a high degree of uncertainty.
As you can see in the screenshot to the left, I was effectively estimating where along the curb I believed I was and then selecting the appropriate layer and adding the feature, with the uncertainty level for the device well outside the range (49.5 meters in this case) I’d set as acceptable (1 meter). I waited for a few minutes at the start of the first curb to see if the device would recalibrate, but GPS drift amongst buildings in a dense urban environment proved to be too great to ever calibrate to an acceptable level.
The Coord Collector app uses augmented reality (AR) to determine the relative distance that the iOS device has traveled since the beginning of any curb. The major benefit of this, as compared to Collector for ArcGIS, is that the surveyor’s distance is being recorded automatically; they don’t need to determine where they are on the map and select individual points, as they do in Collector for ArcGIS, to add assets and features to the map.
Incorporating a measuring wheel with the Collector for ArcGIS increased my confidence in the accuracy considerably, although on average, the location of features collected in Collector for ArcGIS were about 1.5 m off from those collected in Coord Collector, with some features 3 or even 4 meters apart.
Part of the confidence in the accuracy of Coord Collector’s data comes from Coord’s requirement of having two independent surveys of the same area for a curb to be considered “complete” - for this, we had one of our other employees go out and capture data of the same curbs to validate my collection.
Post-collection: data editing and regulation building
Although Coord Collector was faster than Collector for ArcGIS in just collecting and labeling curb data, the gap in time savings between the various approaches actually widens in post-collection processing. Collector for ArcGIS data shows up as points and lines, but related to just the assets - to convert it to regulations, as we do in Coord, and make it usable for measuring total curb capacity, you have to understand what the underlying regulations are for each sign type, based on the text and symbology, and then create a regulations layer in ArcGIS with these competing priorities appropriately setup.
Coord’s system takes these assets and automatically extracts the writing and symbols on the signage within a few minutes. Continuing without user intervention, it then passes these assets through Coord’s regulations engine, resulting in a complete set of regulations, and curb feet allocated to each type of regulation, in mere minutes. I exported the data from Coord’s analytics page and imported it to ArcGIS for this test, just to demonstrate the compatibility of the system and interoperability.
For the data collected with Collector for ArcGIS, I explored incorporating some basic image recognition and translation, including passing it through free online OCR systems. There are a number of paid services that also provide OCR of course, such as Amazon’s Rekognition. Ultimately, for the quantity of data collected in this test, this route wasn’t necessary, but for a larger collection area, this would be a necessity.
Once I’d done this, I referred to the NYC parking guide to determine what the official rules were based on what I’d collected. I then created a separate layer for regulations and added the regulations, as best as I could off of lat/lons and priorities.
A limitation of ArcGIS is that representing change over time, which many regulations do, is complicated. Coord has pre-built filters for time of day and day of week, along with a few others related to curb usage. While I could represent the parking regulations for this set of curbs in ArcGIS by including the entire range of time for the regulations, this doesn’t allow me to understand how curb capacity changes over time, especially as I add in additional blocks.
Here were the final results:
------ Collection Methods ------
|Collection & Translation Steps||Coord||ArcGIS||ArcGIS with measuring wheel||"x" times faster than ArcGIS with measuring wheel|
|3 min||20 min||None - Already setup from ArcGIS collection||6.7|
|Collecting One NYC Block||5.25 min||10.5 min||16 min||3.1|
|Editing the data||0 min||10 min||6 min||N/A - no time spent editing data in Coord|
|Importing and styling in ArcGIS||5 min||30 min||30 min||6|
|Regulation translation||2 min||60 min||60 min||30|
|Total Time||15.25 min||130.5 min||112 min||7.4|
The Coord Collector app has been built and optimized for exactly the type of collection and data processing required to build parking regulations, and other curb-focused rules, out of asset data. Collector for ArcGIS is a more multi-use tool, designed to be used by everyone from foresters trying to understand the spread of certain diseases amongst a tree population to road race managers keeping track of the number and type of emergency vehicles present along the race course.
Collector for ArcGIS does allow you to integrate with professional grade GPS receivers that can get accuracy to the centimeter, but this is a costly investment on top of an iOS device and considerably increases the collection time.
Based on the requirements of most parking studies, the Coord Collector app appears to be the best choice for both speed and flexibility, especially for any urban environments where the tall buildings will impact phone GPS.
Commenting is disabled on this post.