Using Regression Analysis in Transportation Analytics- A Use Case

Recent trends indicate a surge in application of predictive analytics methodologies in the Logistics and transportation domain. Regression analysis, which is a widely used predictive analytics tool, can sometimes also be used to evaluate the key variables that are driving your transportation costs and can provide some really useful insights. Below is an example from a real project that I worked on years ago. I have changed the data for confidentiality purposes but this is an interesting example to illustrate how regression analysis can help you highlight some of the key drivers of your Transportation cost.

We were trying to determine the cause of variation in logistics costs across different locations. We used regression analysis to predict and analyze transportation costs, based on the following transportation cost drivers identified by the company:

  • Quantity shipped
  • Weight shipped
  • Item cube (Dimensions)
  • Shipment miles
  • Allocated fuel cost

We did this analysis for outbound shipments from five DCs but I will use the data for only two DCs to illustrate the kind of insights such an analysis can provide.

Location A

So, based on the shipment data, we ran the regression analysis and we can now analyze the output of the analysis (I am assuming that readers are familiar with interpreting the output of an regression analysis).

Allocated Freight cost = A + .00310Qty Shipped + 0.00306Weight Shipped + 0.00144Item Cube + 0.030347Shipment Miles + 10.8 Allocated Fuel cost

R(adj)=99.4

P Values:

  • Qty Shipped 0.041
  • Weight Shipped 0.000
  • Item Cube 0.669
  • Shipment miles 0.000
  • Allocated Fuel Cost 0.000

So what does the above output tell us ? Obviously Shipment miles and fuel cost are significant drivers but what else? If we look at item cube p value, it tells us that Item cube is not significant. We are probably weighing out the trucks-the products that are being shipped from this DC are heavy items (and when we validated, we did find that they were large, bulky furniture).

Location B

Now let us analyze the output for other location.

Allocated Freight cost = B + .00097Qty Shipped + 0.00684Weight Shipped + 0.1929Item Cube + 0.000574 Shipment Miles + 8.69 Allocated Fuel cost

R(adj)=84.3

P Values:

  • Qty Shipped 0.506
  • Weight Shipped 0.009
  • Item Cube 0.000
  • Shipment miles 0.189
  • Allocated Fuel Cost 0.000

So what does the above output tell us ? If we look at item cube p value, it tells us that Item cube is significant driver for this location. We are probably cubing out the trucks-the products that are being shipped from this DC are considerable dimensioned (and when we validated, we did find that to be the case).

As you can see, leveraging Regression analysis can provide really useful insights into your drivers. They can also help you create Logistics product segments i.e create segments of products that share the same logistics characteristics.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s