Which mapping stack do you usually choose?
That depends on the use case. For consumer and public-facing apps we often work with Mapbox GL or Maptiler, thanks to their attractive vector styling and reasonable licence costs. For public-sector projects we usually choose Leaflet on top of OpenStreetMap, or a combination with PDOK tiles, because that is licence-free and interoperable. For organisations that already invest heavily in ESRI, we align with ArcGIS Online or ArcGIS Enterprise. We use Google Maps where recognisability matters more than customisation. We always advise based on your data, target audience and existing contracts, not on what we happen to be familiar with.
How accurate is the location in the app?
With the standard GPS chip in modern smartphones, you typically get three to five metres of accuracy in the open air, and slightly less in urban areas. For field surveys, archaeology or cable and pipe work that is sometimes not enough, so we connect external GNSS receivers such as the Trimble Catalyst, Emlid Reach or a Septentrio Mosaic over Bluetooth. With RTK correction via an NTRIP stream, you then achieve centimetre precision. The app automatically falls back to the built-in GPS if the external receiver is unavailable.
Does the app really work offline in an area without a network?
Yes, that is a core requirement for most of the geo apps we build. We download tiles and feature data to the device in advance, store all work orders, photos and measurements locally in a SQLite or MMKV store, and synchronise via a queue once the network is back. Conflict resolution for objects edited by several field workers at the same time is built into the design by default. We test this every sprint in aeroplane mode and on poor 4G, not just on a fast Wi-Fi connection in the office.
Can you integrate with our GIS environment?
In practice, almost always. For ESRI: ArcGIS Online via REST services or ArcGIS Enterprise. For open-source stacks: GeoServer with WMS, WFS or WMTS, or a direct PostGIS database. For QGIS publications: an intermediate step via a WFS feed or an exported GeoPackage. For PDOK data: the open WMS and WFS services. We set up two-way integration where that makes sense, so field captures land in your office GIS automatically and vice versa.
What about GDPR when it comes to people's location data?
A person's location falls under special category personal data under the GDPR, particularly when it is collected continuously or in real time. We build consent flows, data minimisation and retention periods directly into the app, with an audit log recording when and why location is collected. For fleet applications involving personally assigned vehicles, we work with works council agreements and privacy-by-design principles. For projects involving the location of patients, pupils or residents, we carry out a DPIA as standard. We work with your privacy officer or data protection officer to clarify the legal boundaries.
How do you combine a geo-app with AI or image recognition?
Organisations increasingly want a photo taken in the field to be automatically categorised or analysed. For example: an ecologist photographs a plant and immediately gets the species back; an inspector records waste and has the type detected automatically; a property surveyor photographs damage and receives an initial assessment. We build this as a separate AI layer on top of the geo-app, with models running on-device or in the cloud depending on privacy requirements and data volume. For our approach and capabilities, see our page on
AI development.
What is INSPIRE and do we need to comply with it?
INSPIRE is a European directive that obliges public authorities to publish certain spatial datasets in a standardised format and with standardised metadata. If you are a municipality, water board, province or environmental service working with public spatial data, this usually applies to you. We make sure your app's data export produces the correct GML or GeoJSON formats, with metadata in the ISO 19115 format. For private parties, INSPIRE usually does not apply, but interoperability with PDOK datasets is often still desirable.
Can you do indoor positioning without GPS?
Yes. We place Bluetooth beacons (from providers such as Estimote or Kontakt.io) at fixed positions, and triangulation gives a position accurate to within a few metres. For larger buildings we use Wi-Fi fingerprinting on top of existing access points. For BIM-driven routing we load an IFC model into the app, so users can navigate through a 3D representation of the building. Indoor positioning works best for organisations with a fixed physical infrastructure, such as warehouses, hospitals, stadiums and car parks.
Native or cross-platform?
For geo apps we typically choose Flutter or React Native, because the mapping libraries (Mapbox GL Native, MapLibre, Leaflet via a webview) work similarly on both platforms and you keep a single codebase. For projects with very demanding graphical requirements or highly specific hardware integrations, such as deep integration with an external GNSS receiver via a proprietary SDK, we sometimes opt for native iOS (Swift with MapKit and CoreLocation) and native Android (Kotlin with the Maps SDK). We discuss during discovery what suits your requirements and existing engineering capacity.
Who manages the tile licences and hosting costs?
This forms part of the management contract if you take one out. Mapbox, Maptiler and Google Maps use a usage-based model (per map load or per tile request), while ESRI works with named users. We monitor usage, scale the plan up or down, and provide you with monthly insight. For public-sector projects we use PDOK tiles wherever possible, as these are licence-free, with a commercial provider as a fallback for specific styles or regions.
Do you also work alongside our own GIS analysts or developers?
Yes, that is more the rule than the exception with geo projects. Many clients have a GIS team that handles the data layer and analyses, while we build the mobile layer and, where needed, the intermediate APIs. We work in the same Git repository, align with your projections and data models, and pair-program where that speeds up knowledge transfer. At the end of the engagement your own team can maintain the codebase fully independently, with no lock-in to us as a supplier.
