Previously, you may have identified datasets which are critical to your project, but that you don't have and you can't acquire them from another organization. So let's think creatively about how you could capture them and then leverage those assets to fuel your project and create that long-lasting competitive advantage. Going back to our Retail Transformation Project, we need a dataset about customer style preferences and product interactions. We can collect this data before building the product using a well-designed, immersive, ad hoc questionnaire delivered online or through a kiosk in our stores. This is also where our quick wins and developments can create value as the way customers handle these in-store display boxes is indicative of future usage patterns for the style capsule. They could be captured with sales assistants documenting their observations or through in-store video recordings monitored by behavior specialists. The remaining datasets require customer input after they've actually received the product. A quick win might be to place a feedback form on your app or website and offer a discount to incentivize users to respond. But as I said earlier, we need to think creatively. What if you use the superpower of perception to capture data more effectively and to gain richer insights with less hassle for the user? Check this out. Smart sensors could be embedded in the box to monitor if an item was picked up. Now, you may have seen similar sensors in hotels that monitor interaction with products like their sparkling water or the chocolate bars that are placed in your room for purchase. If the item is not returned in the box within a three-day period for example, the information automatically collected by the sensors could then be used to automatically generate the corresponding customer's invoice, streamlining your processes. If you're looking for additional insights, you might put sensors on the items themselves and learn more about how they're handled by the distance each item travels, and for how long. This could provide information on how well the item was received, and this data can be fed into your future recommendations to the customer. You can also add a device that will connect with Google Assistant, which could then chat with customers and ask them questions while they look through the products. There are a lot of new techniques available to capture and share data, including sensors and natural language assistants. The IoT or Internet of Things can play a huge role in capturing the datasets that you need. How about connecting your style capsule to Nest thermostats, for instance, to better understand dressing habits in high or low temperatures based on the items selected at a different time of day? This is usually done through application programming interfaces or APIs, which are simple methods and tools to connect various applications. Think of them as ambassadors between programs from different companies or project teams, which enable them to interact efficiently without deep knowledge or expertise. APIs play a very important role in the creation of your solid data ecosystem as they connect different systems and then make it easier for you to aggregate datasets at each step of your transformation journey. There are several APIs and each serve a different function. In this diagram for example, the API connecting the information from the style capsule sensors to the cloud is called Cloud IoT Core. It's also important to consider at this stage, how you're going to securely store the data that you capture. This is where cloud storage can help you, as we discussed in an earlier part of the course. Google Cloud's built-in defense layers can help you keep your customer's data more private and more secure. A final thing to consider when you're planning what data to capture is what else this data could be used for. When you're capturing data for the first time, there'll be an initial upfront investment. There will probably be maintenance costs associated with it too. Naturally, the investment needs to be worth it. In our retail example, the behavior patterns you observe might also provide useful insights for the merchandising team who designs the layout of your stores. Information from the sensors will allow your customer service team to respond with a more personalized approach to customers' queries. There are several other use cases for this data as you might imagine. Understanding where else the datasets you've identified could be used will help you add further value to your business case, and that's what we're going to discuss next.