From the course: Leveraging Cloud-Based Machine Learning on Google Cloud Platform: Real World Applications

Case study: International Drone Inc. - Google Cloud Tutorial

From the course: Leveraging Cloud-Based Machine Learning on Google Cloud Platform: Real World Applications

Start my 1-month free trial

Case study: International Drone Inc.

“

- [Instructor] So let's take a look at a case study that leverages machine learning. We'll call our company International Drone Incorporated. A fictitious company. And this company sells drones, in essence, resells drones to the agriculture community. The ability to leverage drones to fly over fields to determine things such as hydration, and health of the crop. So keep in mind we have to deal with requirements. We have to deal with the input, we have to deal with building the model, and we have to deal with the output. Need to determine what these three things should be. So the input in this case would be inventory data. The amount of drones that we have available to sell. As well as sales data, the number of drones that we're able to sell. The model, we're able to determine patterns of success and train the model based on what looks like is becoming more efficient for the business. As well as patterns or failures. Where sales and inventory seem to be failing each other. There's too much inventory to sell, there's not enough inventory to sell. Then the output. That includes the diagnostics as to what's truly going on and the ability to understand the likely outcome. So the ability to adjust our ability to keep inventory on hand to enhance sales. To make sure that we have enough drones in inventory to meet the demand in the marketplace. So the solution is machine learning. So we're able to ingest data, in this case from the inventory system, and the sales system, we're able to use unsupervised machine learning to discern from the patterns as to what's going on in terms of our ability to keep inventory on hand to support sales. The ability to automate what's going on. The ability to make sure that we're leveraging this within the business applications already on hand. And ultimately, the ability to find out where we're going wrong and where we're going right. And the ability to try to repeat where we're going right over and over again. And the ability to push that information back into the business systems so we're able to act upon it. So the result is ultimately early detection of inventory issues. We know when inventory is likely going to be depleted based on different patterns that may have been historical data that have been used to train the model. We're able to deal with outcome-based learning and we're able to do this ongoing. So as we process the system and as we run the business, the inventory optimization model continuously improves. Then as a result, we're able to reduce costs because we don't have to keep excess inventory on hand that may be depreciating or becoming outdated. And yet we're going to have enough inventory to meet the demands of the sales that the customers are making.

Contents