As a service provider for IT systems and equipment supplier for vehicles in the waste management industry, c-trace makes a decisive contribution to sustainable waste management in municipalities and cities. Among other things, the collected waste quantities as well as the routes of waste collection vehicles are recorded and used in current products for optimal route planning, early emptying at waste collection points and overall reduction of driven kilometers. In this way, c-trace contributes directly to saving CO2 emissions in the urban environment. A central challenge lies in the transfer of the results of the use of such track-and-trace systems for refuse collection vehicles from one municipality to another. For such a transfer of results, it is necessary to develop and use public data sources that allow a data-based modeling of city districts and street clusters. The range of possible data extends from shapefiles of the street topology to standard land values and demographic and social factors. In a follow-up project, this new database is to be transferred into a new data product in two steps, together with appropriately processed telemetry data.

Making cities and municipalities comparable

In order to be able to compare different cities and municipalities with each other, it is necessary, on the one hand, to determine characteristics that are valid across the board and, on the other hand, to find reliable data sources for these characteristics. 

The aim of the project is to create the basis for the data and ML product described above, i.e. ML models for the result transfer of CO2 savings for municipalities and waste management companies, at the conceptual level. For this purpose, suitable external data sources have to be identified and processed, so that municipalities and neighborhoods not served so far can be accessed. Relevant features are then to be developed on the basis of this compiled database, which can be used in a follow-up project to forecast CO2 savings in new municipalities.

.

Summary and outlook

In the transfer voucher A project "QuaReduk" the definition of important similarity features for cities with their external sources could be successfully worked out. Furthermore, it could be shown how an automatic extraction of the data can be done. This was an important preliminary work for subsequent projects. The features from external sources developed here will be used in the future to classify previously unknown municipalities and neighborhoods in terms of data and to put them into a relationship with the areas located in the customer base of c-trace. 



To this end, cluster models will be generated from the external data in subsequent projects. In a second step, regression models - or similar models based on other machine learning (ML) methods - will be created to enable the transfer of results between urban areas. This would then make it possible to create a new, data-based product that can be used for internal sales purposes, but also for consulting services for municipalities and waste management companies. In addition to the development of the ML product, it is also conceivable to integrate further features in the future and to update the existing data. This concerns, for example, data in the regional database on Census 2011. More recent results from Census 2022 will be added to the database. Therefore, an update of the feature sources in a project at the end of the year is also conceivable.

.