What It Does
This Project Type in StreetLight InSight® analyzes the aggregated home and work places of visitors to your selected locations and the share of visitors who are tourists. The Metrics are derived from our Location-Based Services data source, which we recommend for describing personal travel behavior.
With the Visitor Home and Work Analysis, these Metrics are more customizable in the following ways:
- Customizable Day Parts: You can define up to 24 time periods to study, plus an all-day average time period. For example, if you’re interested in the work locations of lunchtime visitors to a restaurant, you can zero in on the lunch hour that you define. It could be 12 pm to 2 pm, 12 pm - 1 pm, or 1 pm - 2 pm - and you could even study all three hour-long periods separately.
- Customizable Day Types: You can create your own definitions for a type of day. For example, you can include Friday as part of the “weekend,” or define “weekdays” as only Tuesday, Wednesday, and Thursday. This can be particularly useful for analyzing tourist hotspots.
- Customizable Data Periods: You can choose the specific time periods to analyze, including specific dates. This can be useful when studying trends over time, such as seasonal analyses.
The large size and representativeness of our Location-Based Services data allows us to provide more granular insights for home and work locations than we could previously provide – while still protecting consumer privacy. And whether you’re an urban planner or travel demand modeler, analyzing the home and work locations of groups of people that visit locations at specific times of day helps you make better data-driven decisions.
To give you a sense of how these Metrics work, we’ve run the likely home locations of visitors to the Denver Tech Center from 6 am – 10 pm on weekdays (see below).
The Denver Tech Center is the Zone outlined in black above. The 1km grid represents the likely home locations (in aggregate) of Tech Center visitors who visit the building on weekdays between 6 am and 10 pm.
As you can see, the majority of visitors’ home locations are clustered near the Tech Center. However, there are also hotspots in Denver’s Lodo and Capitol Hill neighborhoods, as well as in Aurora’s suburban Sable Ridge and Willow Park neighborhoods. Because the Denver Tech Center covers more than 850 acres and houses some of the largest employers in Denver, we expected to see this fairly large spread of home locations.
In addition to home and work locations broken into 1 km grids, you’ll be able to analyze a variety of other Metrics with the Visitor Home and Work Analysis Project. Project downloads will include .CSV files with details on the following Metrics:
- Visitor Activity
- Tourist Summary
- Visitor Demographics including Education, Family Status, Income & Race
- Visitor Home Grids, Blockgroups, ZIP Codes, States and Metro Areas
- Visitor Home Distance
- Visitor Work Grids and Blockgroups
- Visitor Work Distance
- Local Demographics including Education, Family Status, Income & Race
How It Works
Determining the Home and Work Locations for a Device
First, we look at devices' locations during nighttime hours, when people tend to be near their residences. We assign a probability that devices are affiliated with a particular census block based on how many nighttime hours they spend there. A device is disaggregated, so 30% of it might belong to one block, 30% to another, and 40% to a third. For clarity’s sake: we do not have any personally identifiable data (just points in space and time) about the devices' owners.
For the Visitor Home and Work Analysis Project, we also create a reference grid across the US and Canada. A 1 km grid is chosen as a home grid if it is one of the top 5 locations for the month where the device spends nights. In the US this is restricted to where people can live as determined by residential blocks in the US. In Canada, the homes are where the device spends the nights with no other restrictions. A grid is chosen as a work grid if it is one of the top 5 locations for the month where the device spends the weekdays. We then weight each possible home location based on the number of nights the device spends there throughout the month.
Nighttime hours are defined as 7 pm to 8 am (+/- 1 or 2 hours depending on the exact time zone and data month). Work hours are defined as 11 am to 4 pm on weekdays (M-F) (+/- 1 or 2 hours depending on the exact time zone and data month). This means that if someone is working the night shift at a factory in a grid that overlaps a block where people reside, the factory grid may appear as the device's home, while the grid where the night-shift worker sleeps during the day may appear as the device's work location.
Defining a Visitor
Anyone who’s device “pings” within the Zone of analysis will be considered a Visitor to your Zone. A Visitor could be someone who spends extended time in your zone, or who just passes through long enough for a device to “ping”. However, Visitors will be weighted differently depending on whether they live or work inside your zone of analysis.
Visitors living and working outside of your Zone of analysis will be given full weight. For devices that live within the Zone of analysis (permanently or seasonally), we still count them as having visited the Zone; however, we weight them less than other Visitors. Devices that live within the Zone of analysis are given 1/3 the weight of a device that does not live in the Zone. This is because we expect them to be frequently in the Zone of analysis, and thus want to give more weight to those who are true "Visitors" to the Zone.
For a seasonal resident, it will depend on where the device has spent the calendar month. We define home locations by examining the top 5 locations a device spends nights throughout a calendar month, and then weight them accordingly. This means if a device spends 3 weeks in Boston, and 1 week in Miami, we will allocate 3/4 of its home location to Boston, and 1/4 of its home location to Miami.
Once we determine home locations and allocate visitor weights, we then continue our normalization process. For LBS data, we perform a population-level normalization for each month of data in order to determine a "pop-factor". For each census block, StreetLight InSight measures the number of devices in that sample that appear to live in that block and makes a ratio to the total population that are reported to live there according to the 2010 US Census. A device from a census block that has 1,000 residents and 2 StreetLight devices will be scaled differently than a device from a census block that has 1,000 residents and 200 StreetLight devices. From there, we calculate a single "pop-factor" for each device, by adding up the assigned population from each of its census tracts, weighted appropriately.
Since residential blocks are highly affiliated with income, race, and other demographic features, this approach normalizes for bias of those features.
Frequently Asked Questions
How big should my Zones be?
We generally recommend that your Zones of analysis are no larger than 4 square km. If you need to analyze an area larger than 4 square km, consider breaking your Zones into smaller sections, or reach out to email@example.com for assistance.
How many Zones can I analyze in one Project?
When running a Visitor Home and Work Analysis, you can analyze a single Zone or a larger Zone Set composed of multiple Zones. Each Zone analyzed will contain its own row(s) of data in your Metrics. Keep in mind that including many Zones in a single Project will increase Project processing time.
If my Zone covers a highway or another road, will those trips get counted in my Project?
The Visitor Home and Work Analysis will take into account any device that “pings” within your Zone of analysis. This may include someone stopping in your Zone and staying there for an extended period of time, or someone walking, biking or driving through your Zone. If your Zone covers a road or highway, it’s possible that some of those trips will contribute to your Metrics.
Does the Visitor Activity Index represent an estimated visitor count?
The Visitor Activity Index is a measure of the relative volume of visitors to the Zones. The values are provided on an index and do not indicate the exact number of visitors. Values can be compared to other Zones in the same Project, or to Zones in other Visitor Projects, to understand how the relative volumes of visitors at different Zones match up. Values cannot be compared to other StreetLight indexes such as the StreetLight Trip Index.
If you would like to expand your Visitor Activity Index values so that they resemble estimated visitor counts, this can be done by comparing the Visitor Activity Metrics to real world counts, either for one of your Zones of analysis, or a nearby location. Say you had a daily count of visitors to a nearby Visitors Center on a Saturday. You can compare this count to your Visitor Activity Index values for a Saturday and create a ratio or multiplier value that can be used across other Zones in the Project to estimate total visitors to each Zone.
How is Visitor Home and Work Sample Size different from the Visitor Activity Index?
In your downloaded Metrics, there is a file called "your_project_name_sample_size," that contains an approximate device count. This device count represents the total unique devices used in your Visitor Home and Work Analysis. Unlike the Visitor Activity Index, this not a normalized volume but rather a single unique device value that spans all Zones, day types, and dayparts used in a Project. Neither values represent an exact number of visitors and are not comparable.
The Visitor Home and Work sample size is not comparable to sample sizes provided for other Project Types. Learn more about how we determine sample size for our Metrics.
How are international visitors handled?
Since our LBS data is app-based, rather than cell carrier based, we have the potential to capture international tourists traveling within the US in our sample as long as they use one of the apps aggregated by our data supplier, Cuebiq.
One caveat with international devices is that we aren't able to detect their "home location" outside of North America. So for example, if an international tourist spent a week in Los Angeles before coming to the Bay Area, then we might attribute their home location to Los Angeles. Or if they spent half their time in Los Angeles and half their time in San Francisco, we might attribute their home location 50% to LA and 50% to San Francisco (we allow a device to have up to 5 potential home locations based on where they've spent time across a calendar month). Ultimately we aren't able to separate international from national devices in our sample.