Skip to content

Data Engineering

From analyzing historical data to streaming real-time data to machine learning predictions, we guide our customers every step of the way along the data journey, helping them to integrate, organize, and analyze all of their data in the best possible way.

We place a special focus on streaming data, such as telemetry or location data, which must be processed in real time. We also take into account the special challenges of location-based data, which can occur in the areas of mobility, logistics or customer data, for example. 

By using managed products from the Google Cloud Platform, we ensure high-performance and stable data pipelines and data stacks that require little effort to operate. Based on Google Maps Platform, Google Cloud Platform and Looker, we are able to implement individual solutions quickly and with high standards and to develop lightweight applications. 

Our services include extracting data from various sources, data processing, aggregating the data and finally visualizing the data in web-based applications. We use products such as geo-databases, BigQuery, Dataflow and Pub/Sub for this purpose.

Our services:

Data management backends

We develop customized backends for geographic data that enable the display of simple points as well as complex geometries on the map. We implement processes that are adapted to individual customer requirements. We rely on serverless development based on a micro-services architecture with interchangeable components. The backends are hosted in the cloud and help to structure, manage and increase the quality of the data.

Approx. 80,000 merchant data worldwide are stored and processed in the backend

Geo Data Quality Tools for Conti

For our customer Conti, we developed a backend for an international dealer locator. The requirements included a decentralized system where address data must be in a specific format. We built a backend that included geocoding of address data for 80,000 dealers using Google Geocoding API. We also cost-effectively integrated address validation into the application's backend processes.

address validation for Continental

Data Streaming

Data streaming refers to the real-time processing of highly dynamic data such as tracking and sensor data to evaluate a current state. We develop processes to transmit, process and aggregate the data. 

The streaming is done in the cloud. While the data is displayed in the frontend, automatic processes run in the background, such as sending messages at certain thresholds. Our team supports you in setting up these processes, in data engineering on the cloud and in developing ETL pipelines.

In a joint research project with HVCC, we analyzed the movements of barges and visualized the data in Looker Dashboards.

Real-Time Tracking & Operational Dashboards

For the Hamburg Vessel Coordination Center (HVCC), we developed a cloud-based data analytics concept from data preparation to analysis and visualized the results in Looker dashboards and Kepler.

The data basis was AIS streaming data stored since early 2021 from our AIS Cloud Data Streaming research project. HVCC provided GPS data from lighters, which we supplemented with open weather and tidal data. The AIS Dashboard visualizes the results from data collection and asset tracking using data streaming.

AIS Dashboard

Machine Learning

Machine Learning offers companies a wide range of possibilities to automate processes, make predictions based on historical data and optimize business processes. 

With our expertise in Location Data & Cloud, our services in the area of Machine Learning relate in particular to the topics of Demand Prediction, Image Recognition and ETA calculation. We rely on cloud products such as BigQuery ML and Vision ML in the cloud and APIs such as the Google Directions API. Our solutions enable our customers to understand complex relationships and make data-based decisions.

ETA calculation

When calculating ETA (Estimated time of Arrival), we perform route calculations depending on traffic conditions and asset location. We work with large amounts of route and address data to calculate travel time polygons and derive which point data is in the area. For example, we developed an accessibility tool for the Dusseldorf fire department.

Location Tool fire brigade

Demand Prediction

Predictions can also be derived in the area of mobility based on historical data. For a sharing provider, we created a Data Stream Pipeline to process, store and analyze real-time data from mobility sharing providers worldwide. The processed data provides interesting insights about hotspots and typical routes at different times of the day, days of the week, and seasons. Using Google Datastudio, Looker and Customized Frontends, we can visualize and illustrate the results.

data analytics for sharing providers

Image Recognition

Use cases in Image Recognition include, for example, inferring weather conditions from traffic cameras or classifying satellite data to control illegal logging or crop cultivation. When classifying satellite data, we train an algorithm on what kind of building is shown or what kind of grain is in the image.

Your experts for Location Data & Cloud:

  • We are experts in geospatial data and answer questions that cannot be answered with standard solutions.
  • We are very familiar with the common providers of geo APIs (Google, Mapbox, Targomo, here, Graphhopper, etc.) and know how to best aggregate and analyze geodata.
  • We have over 15 years of expertise in dealing with particularly large data sets and dynamic location data
  • We support our customers as sparring partners and consultants, create a good understanding of the technical basis and quickly work our way into complex issues.
  • We use a holistic approach and think application from the end: only when we know what is to be achieved with the presentation of the data do we make sound decisions for data preparation and presentation.
  • We focus on Google's Data Cloud and therefore bring a very deep knowledge in the relevant products such as Dataflow, BigQuery and Dataproc. We can apply the cloud technologies appropriately to the (geo) data projects.