Driving Times 2.0 – An update
We wrote about our view on driving times and how we think that the analysis of these driving times must/will change in our February blog. Today we provide you an update of the project:
Consultation about influential factors
In the past month, we’ve contacted public transport specialists about the factors that influence driving times. We’ve received many valuable responses and would hereby like to thank everybody; you will all receive a complete list when the time comes and an overview of how we’ve incorporated the data into TMWalker.
Based on this list of influential factors, we are now searching for the correct data. This could be public data, such as the weather, that would be available for all TMWalker clients, or company-specific data with which one can improve their own TMWalker environment. Think; messages about the communication with traffic light installations and vehicle log files. As a client, you decide with whom to share this company-specific data.
Back-end: tool selection for loading the staging area
eXtractorONE showed up as the best tool for extracting pubic data sources into our staging databases. eXtractorONE generates a database and inserts this data in no time; whether it be four thousand CBS reports or a NeTEx timetable.
Back-end: iron and steel
Since the data we use is not all public data, we chose to stick to our own servers. This way we can guarantee security and host location. We are currently holding ‘the beast’ (as we call it lovingly); a computer system with 22 processors and 1.5TB of RAM. Enough capacity to host complex algorithms and we are very much looking forward to using it.
Back-end: business rules
Both the algorithms we use to calculate driving times among other performance indicators, and the output of all the TMWalker analyses are currently undergoing change. We are doing this as preparation so that we are able to add context information at a later stage. Additionally, we use this ‘sweep’ to generate automated documentation about how the calculation was actually executed. This is becoming more and more important as there are multiple factors that can influence this calculation, obvious or not. (Is the calculated driving time due to a snowy day, or are snowy days excluded from the calculation?)
We are working away on the front-end part of this project too, together with freelance Service Designer Romee Noorman. We are drawing up a user experience that matches our target group (you). We expect to be able to start testing at the end of this summer, and are thus looking for a select group of test users. Are you curious, or do you want to determine how happy our happy flow makes you yourself? Please send us a message (email@example.com) and join the bèta test group.
Summarized: lots of changes, lots of improvements. More information next month.