Interesting discussion, Benje. We did some work a few years back tracking vehicles in Victoria, Australia with a plan to calculate fuel consumption and emissions based on inputs such as vehicle year, make, speed, topography, acceleration etc. It was a lot of data and even then we lacked important inputs (that influence efficiency) such as the condition of the engine and the tyre pressure). There are always limits to this. However, generally speaking, we found efficiency gains would come from just a few standard improvements, which meant after collecting data for a while, you didn't need any more. In your case, you're looking for individualised improvements based on routing and so forth, so that's another level of complexity again. However, maybe this work would have the same outcome: that is, a few general tips for improvement help most drivers.
You're right, it is another layer of complexity to get vehicle-level data. That's why it can't be just a smartphone app solution. In the research in Stellenbosch, I think they used a device that plugs into an OBD port. (And only if it's a newer vehicle.)
>a few general tips for improvement help most drivers
Maybe. Or maybe more.
If we treat informal transport providers as SMEs (which they are), then this is critical business operating data. As the adage(s) goes, what gets measured matters, and what gets measured gets changed. Daily and weekly data would go towards continuous business improvement.
As I understand it, some of the work done in this level of data collection with a taxi union in Cape Town allowed the union to reduce drivers' hours to just 9 hours a day and still meet demand. It allowed them to get a day off once a week while moving them to a steady daily salary (instead of a daily take of whatever is left from fare revenues minus vehicle rent).
It was a study of private vehicles commissioned by an automobile club. They used a widget plugged into the car somewhere to collect the data, but even then, we had to make broad assumptions about vehicle condition and so forth.
Interesting discussion, Benje. We did some work a few years back tracking vehicles in Victoria, Australia with a plan to calculate fuel consumption and emissions based on inputs such as vehicle year, make, speed, topography, acceleration etc. It was a lot of data and even then we lacked important inputs (that influence efficiency) such as the condition of the engine and the tyre pressure). There are always limits to this. However, generally speaking, we found efficiency gains would come from just a few standard improvements, which meant after collecting data for a while, you didn't need any more. In your case, you're looking for individualised improvements based on routing and so forth, so that's another level of complexity again. However, maybe this work would have the same outcome: that is, a few general tips for improvement help most drivers.
Kevin
Hey Kevin,
What kind of service were the vehicles providing?
You're right, it is another layer of complexity to get vehicle-level data. That's why it can't be just a smartphone app solution. In the research in Stellenbosch, I think they used a device that plugs into an OBD port. (And only if it's a newer vehicle.)
>a few general tips for improvement help most drivers
Maybe. Or maybe more.
If we treat informal transport providers as SMEs (which they are), then this is critical business operating data. As the adage(s) goes, what gets measured matters, and what gets measured gets changed. Daily and weekly data would go towards continuous business improvement.
As I understand it, some of the work done in this level of data collection with a taxi union in Cape Town allowed the union to reduce drivers' hours to just 9 hours a day and still meet demand. It allowed them to get a day off once a week while moving them to a steady daily salary (instead of a daily take of whatever is left from fare revenues minus vehicle rent).
It was a study of private vehicles commissioned by an automobile club. They used a widget plugged into the car somewhere to collect the data, but even then, we had to make broad assumptions about vehicle condition and so forth.