Improving location data for accurate algorithms and powerful visualisations
Case Study Summary
Website: margemonitor.nl
Industry: Business Intelligence consultant to logistic companies
Impact made:
- Correct coordinates enables correct and powerful visualisations on a map
- To correctly aggregate data based on location data you need similar ways of writing for example city names
- To calculate profitablity in logistics, coordinates need to be correct to be able to determine the distance between loading and unloading points
Challenge
The location data of the orders can be input by the end customer or input manually by the planner. This introduces the risk of incorrect data, inconsitent data or incomplete data. This hinders analyzing the delivery after the fact.
-
Profitability calculation
To calculate the profitability of a delivery, we need to accurately determine how much time is spent on each delivery. This is especially hard to do if you are distrubiting to multiple customers in the same truck. To correctly share the time spend driving for each delivery you need correct loading and unloading coordinates. -
Powerful visualisation
Showing profitablity of certain loading and unloading places on a map is super powerful. This can show certain spots that might be outliers in your network or show that certain prices for your services are not enough to deliver the goods in a profitable way. -
Data quality issues
If we want to show how much turnover or profit a certain city provides to the company, inconsistent ways of writing the city name hinders this. Inconsistent spelling of city names makes grouping and filtering data unreliable. For example, filtering on the city name “Rotterdam” may exclude records entered as “rotterdam”, leading to incomplete or incorrect analyses.
Solution
To improve the location data I used the HereMaps Geocode API. This API I can call within a Qlik Cloud environment with the locations data that needs improving. I input the adress, postcode, city and country and as a result I get location data that contains coordinates, correctly formatted location data and a measure how confident HereMaps is about the result returned. Missing data in the input can cause no result or not an accurate enough result, to combat this i wrote a fallback query that contains the postcode, city and country. If this fails, the fallback is the orginal coordinates and location data.
In this method 99+% of location data inputted has a good result.
Results & Impact
- Algorithms to determine profitablity have the right input coordinates
- Visualisation of the data is now possible on maps
- Location data is formatted in the same way preventing analysis mistakes
Dashboard & Reporting Showcase
Note on Data Privacy
All sensitive information such as customer names, employee details, and specific financial figures have been censored in these screenshots to protect business confidentiality while demonstrating the solution's capabilities.
Visualisation showing turnover and profitablity
-
Let's have a virtual coffee together!
Want to see if we're a match? Let's have a chat and find out. Schedule a free 30-minute strategy session to discuss your AI challenges and explore how we can work together.