The client requested a comprehensive analysis of its transported items within the United Kingdom and Ireland that were GPS tracked. Therein, they wanted to know where their items have been transported and mainly where there have been located in order to find those that deviated from their stores and where they finally ended up.
How we handled this
The dataset was provided by the client consisting of an unique device ID, timestamp, longitude and latitude. In addition, another dataset was provided of all their stores with addresses but without the longitude and latitude data. With this data the following steps were performed in accomplishing the requested goal:
- Initially, the requirements were discussed on the purpose of the project
- The data was explored, cleaned and data engineered. In addition, some manual work was performed to find the longitude and latitude data based on the store addresses provided
- The coding script was designed to be incremental future proof usage that would enable new incoming data to be used for further analysis
- A distance list was created of previous calculated distances in km’s and in miles in order to prevent calculating the same distances multiple times.
- Rule-based modeling was used to identify the lost items and those that are divergent
- A well programmed, cost-efficient and structured R script that allows for newly added data to be analyzed.
- A List of divergent item locations
- A clear visual presented world map that identifies all their stores located as well as the tracked items and those that are divergent.