Enhanced Data Coverage with Mapbox Movement V4

New release features broader coverage, enhanced privacy, and increased impact

Sofia Heisler

Feb 14, 2024

Enhanced Data Coverage with Mapbox Movement V4

New release features broader coverage, enhanced privacy, and increased impact


Sofia Heisler


Feb 14, 2024

Mapbox Movement V4 is now available, boasting a major data upgrade that increases data density and data coverage. This release allows us to resolve activity with greater accuracy and consistency, and expand data coverage by 25% in metropolitan areas to over 300% in rural areas. New privacy filtering obscures activity patterns on private land with low activity while maintaining data integrity in public spaces and roads. An improved outlier detection process now has a 95% accuracy rate in identifying and adjusting outliers caused by data artifacts, while avoiding any adjustments to naturally occurring high-activity quadkeys.

All of this means that the mobility data needed to make decisions with confidence can now be used reliably in more places with a cleaner and more accurate view of footfall activity around the world.

Mapbox Movement Data is the world's foremost privacy-first mobility dataset, built upon first-party telemetry data and powered by over 700M monthly active users and 45,000 applications. Mapbox customers use insights derived from this data to select high-value retail locations, analyze risk, forecast market trends, understand global infrastructure demands, and more. The latest release significantly boosts quality while maintaining the privacy standards that we prioritize above all else. 

Enhanced Movement V4 data

Vastly improved activity coverage in urban and rural areas

In previous versions of the Movement dataset, we relied on sampling telemetry probes (timestamped latitude/longitude telemetry coordinates). While this resulted in fairly comprehensive coverage in urban areas, it caused lower coverage in rural areas where signals tend to be more scarce. The latest version of Movement features an updated sampling methodology, maximizing the number of probes captured when calculating activity. This has led to increases in data coverage ranging from 25% in urban areas to 300% or more in rural areas.

Data coverage in the urban area of Bristol, UK and nearby rural areas in V3 dataset (left) versus the latest V4 dataset (right).

More granular localized app adjustments

The telemetry collected by Mapbox is sourced from tens of thousands of mobile applications. When combined, data from these apps allows us to generate a complete picture of driving and non-driving activity throughout the world. However, mobile device adoption and usage patterns vary significantly across urban, suburban, and rural residents. Utilization of individual mobile apps can also go up or down over time as new apps enter the market or older apps become less popular. One of the significant challenges of estimating activity volumes using telemetry is distinguishing the changes in observed activity that are caused by these market shifts from actual "on-the-ground" changes in mobility patterns.

To do so, Mapbox has developed a proprietary source mix correction algorithm that automatically adjusts observed activity for changes in the composition of the mix of mobile apps from which the data is sourced. 

The V4 data release  introduces several changes to our source mix adjustments:

  • Higher spatial granularity of source mix corrections: Starting with data for August 2023, changes in the mix of source apps are adjusted at a significantly higher granularity than before (a granularity of less than 1 square km instead of city-wide). This increased granularity corrects for more localized swings in source data that may have previously been averaged out.
  • Better adjustments of source mix by transport modality: Mobile apps tend to reflect traffic of only one modality (for example, only driving or only walking data). V4 data introduces improved granularity of corrections for app-specific volume changes for each modality, significantly increasing the accuracy of driving and non-driving activity. 
  • Better stability of source mix corrections: Given the enormous volume of applications providing Mapbox with telemetry data, it is not uncommon for occasional bursts of spotty or erroneous raw data to be sent to the algorithm, which risks skewing the source mix corrections. Improved thresholds for the inclusion of sources in the adjustment algorithm now filter out such data and prevent it from influencing the corrections applied to the data.

Improved outlier detection

Despite all of the adjustments described above, activity derived from telemetry can occasionally be prone to outliers that are not reflective of activity on the ground. An example might be a mobile phone store that has a large number of phones on display; all of these phones might be constantly sending us telemetry, creating the illusion that there is significant activity in that location. Movement Data V4 includes a fine-tuned outlier detection module with a more targeted algorithm to find and adjust for outliers, resulting in a 95% accuracy rate in identifying and adjusting outliers while avoiding any adjustments to naturally occurring high-activity quadkeys. This provides customers with a cleaner and more accurate view of where high activity occurs in our datasets, without compromising data integrity.

Two examples of high-activity quadkeys evaluated as potential outliers: on the left, the high-activity quadkey is surrounded by much lower-activity neighbors so it is identified as an outlier and adjusted; on the right, the high-activity quadkey has other high-activity neighbors nearby and is not adjusted.

Privacy filtering

A key feature of all data produced by Mapbox is our commitment to user privacy. For the Movement dataset, this means a delicate balance between ensuring that the movements of individual users of apps running a Mapbox SDK can never be traced using the data and reporting accurate data even in areas where activity may naturally be low. A new, updated algorithm in our privacy filtering module blocks activity patterns on private land that fail to pass minimum activity thresholds, while still maintaining data integrity in public spaces and roadways. These changes allow for more accurate activity reporting.

Built on Boundaries

Movement V4 incorporates the latest release of country boundaries from the Mapbox Boundaries dataset, the leading source for global geopolitical, administrative, postal, and statistical boundaries in the market. The move from Boundaries V3 to Boundaries V4 includes better data coverage and refreshed geometries, summarized here.

Mapbox Movement empowers data-driven decisions

Mapbox customers leverage Movement data to analyze the impact of policy decisions on movement, optimize site selection, and allocate resources strategically. By harnessing high-fidelity Mapbox Boundaries, they gain insights into mobility trends at state, county, and district levels, informing more effective policy decisions.

“The wave of adoption of Mapbox tools has been unprecedented and has given Mapbox a unique capability for massive, decentralized data collection. By making Mapboxdata commercially available, we expect Mapbox to power important use cases such as on-demand delivery and massive route planning, where their 300 million miles of daily data gives them an important competitive advantage." - Brent Iadarola, VP, Mobile & Wireless Communications Frost & Sullivan

Elevate your location intelligence with globally available data (140+ countries) from Mapbox Movement. Download free samples or request a personalized demo today to find out how Mapbox Movement data can power decision-making. 

Heading 1

Heading 2

Heading 3

Heading 4

Heading 5
Heading 6

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.

Block quote

Ordered list

  1. Item 1
  2. Item 2
  3. Item 3

Unordered list

  • Item A
  • Item B
  • Item C

Text link

Bold text




No items found.
No items found.