You need to have a machine learning or app development team tomorrow? And can’t wait for a lengthy team ramp-up? We build code on the first day. While we hire and train an in-house team for you. No time lost. No matter where you are in your life cycle we can support you all the way.
From building a winning data strategy to machine learning prototypes or even full stack prototyping. We are your go-to team for anything mobile, artificial intelligence and location data. We already helped leading brands from concept to production hand-over.
World Class Algorithms in Python, R, Matlab, CNN, Scikit.
Data Lakes in Redshift, Athena, Hadoop.
Custom Analytics and Insights based on Big and Real-Time Data.
Rapid Prototyping in iOS, Android, Cordova, ReactNative.
TASK:
To evaluate the performance of tracking the truck arrival and departure for the purpose of a more efficient use and allocation of parking spaces, an approach needed to be developed that allowed measuring the accuracy of the detection given the special circumstances regarding devices, mode of transportation and driving patterns.
APPROACH:
We developed a test app with mobility detection following the visual guidelines of the client to evaluate performance by comparing the automated detections with manually logged events. Drivers had to give their feedback in different situations with minimal distraction (one click) on their tablet or smartphone devices that are installed in the trucks.
TASK:
The idea was to see how app detected mobility data provides valuable insights into the whereabouts of people throughout the day and identifies hot spots of activity for store planning. Furthermore, it was of interest to see if places of origin can be identified for specified arrival zones where the stores are located.
APPROACH:
We extracted raw data of arrival and departure events (around 4.5 mio) that were detected on phones in the background. After cleaning that data for our purposes, we transformed it to extract additional information about travels and stays by connecting different data points. Finally, we build several visualizations to make that data accessible.
TASK:
For many use cases, it is critical or at least helpful to understand in real-time what mode of transport like a car
APPROACH:
To begin, we collected thousands of trips where users labeled the data, telling us what mode of transport they were using. For these trips, we recorded sensor data for different smartphone sensors like gyroscope or accelerometer. Based on that data we used machine learning to identify patterns and recognize the different modes.
TASK:
To solve the parking problem, the European Commission has invested into a consumer app that guides people to available on-street parking spots. To avoid driver distraction,the system needs to be very simple and intuitive, yet still, provide all the necessary information to find a nearby parking spot.
APPROACH:
In the different countries targeted for the rollout of the solution, we interviewed car drivers about their parking search
