> For the complete documentation index, see [llms.txt](https://docs.jakarto.com/info/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.jakarto.com/info/en/qualite-des-donnees.md).

# Data Quality

Jakarto generates digital twins using mobile mapping units.

These systems collect two types of source data:

* Very high-density georeferenced point clouds
* High-resolution georeferenced panoramic images

These source data are accessible within Jakartowns for viewing, navigation, measurement, and manual asset inventory.

They are also used by Jakarto’s artificial intelligence algorithms to produce certain automated asset inventories and 2D imagery products (raster tiles). Detailed "data sheets" describe the structure of asset inventory data and their respective attributes.

Please take note of the following additional information regarding the inventory data produced by Jakarto:

## Inventory Data Completeness

Only assets visible from the street can be detected. However, visual obstructions during data collection may prevent some assets from being detected.

Our AI algorithms can automatically recognize most assets within a specific inventory (trees, poles, manholes, catch basins, etc.). Our quality control and manual input process increases the completeness of the inventory.

Jakarto does everything possible to provide the most detailed and comprehensive inventories. While perfection is difficult to achieve, the Jakarto team continuously strives to deliver the most complete data possible. The company works tirelessly to ensure maximum accuracy and completeness in its inventories. By using rigorous verification methods, Jakarto aims to provide an outstanding user experience through high-quality inventory data.

## Geolocation Accuracy of Jakarto’s Data

The data collected by Jakarto corresponds to what is visible from the street where the capture units drive. The sensors are located about 2 meters above the ground.

Since the vehicle is in motion during data collection, and since accuracy partly depends on the satellite signals (GNSS constellation) used to determine its location on Earth, the data’s precision varies over time and space.

Accuracy is mostly influenced by the environment around the vehicle. It is better when the vehicle drives through open areas (highways, country roads) and may degrade in obstructed environments (skyscrapers, tunnels, dense canopy). For such complex conditions, our system is equipped with specialized sensors to compensate for GNSS signal loss.

Based on our historical observations, 90% of the time, the 3D accuracy of data collected by the capture system is better than 3 cm (MRSE; equivalent to 5 cm SAS99). For more details, please refer to [the related technical note](/info/en/qualite-des-donnees/note-technique-sur-la-qualite-des-donnees.md).

Accuracy may vary from one capture campaign to another. For example, a fall campaign without leaves may yield more precise data than a summer campaign under dense foliage. In the Jakartowns toolbar, a time switcher option allows users to view different dates when precision is critical.

At all times, accuracy information is displayed in the footer and also available in the 2D map of Jakartowns for every trajectory point.

> **Important:** If the data accuracy falls below Jakarto’s standards, a red banner will appear at the top of the screen to alert you.