Speaker 1: In this video, we're going to turn this boring table of data into this interactive dashboard. Actually, I'm not going to do much at all. I'm going to use AI to do it all for me. And unlike ChatGPT, with this AI tool, it's free to upload files and it creates the dashboard for you. However, it does come with some limitations, so stick around to the end where I cover them and the alternatives. And if you'd like to try it yourself, you can download the example file from the link in the video description. First, let me quickly introduce you to our AI assistant, Claude. Claude is created by Anthropic with backing from Google and Amazon, and it's designed to assist with a wide range of tasks from data analysis to coding and beyond. What's great about Claude is that it's free to use and you can access it from your web browser. Unlike ChatGPT, Claude can generate and display complex content like charts and even entire dashboards in its artifacts window. This preview feature is free to use. You just need to turn it on in your settings, which you'll find by clicking on your initials in the bottom left and then go into Feature Preview. And here, make sure you've toggled Artifacts on. And you can see with Artifacts, you can ask Claude to generate content like code snippets, text documents or website designs. So now that we have Artifacts enabled, let's see Claude create a dashboard for our data. Here I have a table of fictional data by date, region, product and sales channel with units sold, unit price, total revenue, total cost and profit. It's a lot to get your head around in this format. So I'm going to close the file and we're going to drag it in and drop it into the chat. And then in the prompt field, I'm just going to paste my prompt in, which is analyze this data and suggest six key visualizations or metrics that would be most effective for a comprehensive sales performance dashboard and explain why each would be valuable. This way, we're not just getting a dashboard, we're gaining insights into what makes an effective data visualization. So you can see Claude's come back and it suggested total revenue by region bar chart, which will show us which regions are generating the most revenue. We've got sales channel performance pie chart, which is going to compare offline versus online. Then we've got product category performance stacked bar chart, not sure about stacked bar charts, so we'll come back to that. Then we've got monthly revenue trend line chart, which is great for seeing how our revenue tracks over time. We've got profit margin by product scatter plot, which is going to plot units sold against profit margin, revealing which items are most profitable. And then we've got average order value, which we might leave out because we don't have individual orders in our data set. So it can't actually calculate the average order value. So we can see Claude has given us some great suggestions and these visualizations cover different aspects of our sales data, geographic performance, time trends, product analysis, and sales channel split, giving us a comprehensive overview. So now that we have these ideas, let's ask Claude to help us create our dashboard. I'm going to paste in my prompt, which says, based on the visualizations you've suggested, create a visual representation of our complete dashboard with interactive elements, excluding chart number six, which is the average order value and doesn't make sense. Make the line chart span the width of the dashboard. Use the data from the CSV file I uploaded to populate the charts. If you don't say this, Claude might make up its own data. Then we want it to provide this as an HTML CSS artifact that represents how our dashboard should look and use chart.js, which is a JavaScript chart library for creating the charts and make sure to include the necessary JavaScript to process the CSV data and create the visualizations. Let's click send and see what we get. And you can see on the right, Claude is writing the HTML for the webpage to display the dashboard, as well as the CSS for the styling and the JavaScript to generate the charts. And there's our dashboard. So we've got our monthly revenue trend. We've got our revenue by region, our sales channel performance. And if I hover over the different elements in the charts, you can see it gives me the tool tips. It's completely interactive. At the bottom, we've got our product category performance, stacked column chart and profit margin by product scatter chart. So some of the charts are a bit questionable. I probably wouldn't use this stacked column chart. And the scatter chart's not really very effective for this data set. But overall, it's a pretty good first attempt. What is impressive is how Claude has taken our raw data and transformed it into meaningful insights. For example, we can easily see which regions are performing the best, and we can see the breakdown of our sales by online and offline. And down here, we can see which product category performs the best in terms of revenue, cost and profit. Now, I can also click on the publish button at the bottom to publish it, or click the download button to download the HTML file that I can then share with others. And because it's HTML, the file size is super small. So with it downloaded, let me open the file in a new tab. We can see there's our line chart at the top. I can hover over the elements. It's fully functional just as a HTML file. And you can see it's just saved on my C drive. So I could share this file with others over email or in Teams for them to open in their own web browser. Claude has not only created the visualizations, but if I click on the code button, you can see it's written all the code. That is the HTML, CSS and JavaScript code to process our CSV data and populate the charts. Now, if you have JavaScript skills, you can connect this dashboard to a larger data set and have it automatically update the report. But of course, this is just the starting point. Let's see if we can get Claude to up the design quality. So I'll add another prompt asking it to make the dashboard color theme more impressive. And let's click send and see what it comes up with this time. So you can see it's gone for a dark theme. We've got a curved line on our line chart. We've got a donut instead of a pie. We've got a bubble chart indicating the size of the profit margin. And our stacked column chart looks very similar just in the different color scheme. And of course, you could give Claude specific instructions on the color theme. For example, have it in keeping with your brand colors and design best practices to follow. Because although this is eye-catching, the colors don't really aid interpretation as they should. As impressive as Claude's dashboard creation capabilities are, it's important to understand the limitations of using AI for this task. First, there's data volume limit. According to the documentation, Claude can only handle data sets up to 30 megabytes. But in testing, it wouldn't even accept files containing more than a thousand rows and a handful of columns. Second, customization can be challenging. While we can ask Claude to make changes, it doesn't always do what you ask of it. It's not as flexible as creating a dashboard manually, especially when it comes to complex custom visualizations or layouts. And third, there's the data security considerations. When using AI tools, you need to be cautious about uploading sensitive business data. And fourth, integration with other systems is limited. Claude can't directly connect to your company's databases or automatically update with new data. You need JavaScript and Python skills to take the code that Claude gives you and connect it to your source data. And lastly, as with all AI, you must check the output and not blindly rely on its accuracy. So given these limitations, what are some alternatives? One option is to use Excel with AI's assistance. For example, you can use Claude to get ideas and formulas and then implement them yourself in Excel. Another option is to use professional business intelligence tools like Power BI, which offer more flexible and robust features for large datasets and complex visualizations. If you're interested in mastering Power BI, I've got a course designed to quickly take you from novice to expert in creating professional interactive dashboards. In my comprehensive course, you'll learn how to transform raw data into stunning visual reports with more functionality and customization options. You'll discover how to use advanced features, create dynamic charts, and design user-friendly interfaces that make data analysis a breeze. The link to the courses is in the description below and pinned comment. Check it out if you want to take your Excel and data visualization skills to the next level. We've seen how Claude AI can be a powerful tool in streamlining the dashboard creation process, making complex data analysis more accessible. And it's an excellent starting point, especially if you're new to data visualization or you need quick insights. But remember, while AI is a fantastic assistant, there's immense value in understanding the process yourself. And that's why I've prepared something special for you. In this video, I'll show you how to create an interactive Power BI dashboard in just 12 minutes. No AI assistance required. It's a perfect follow-up to what we've learned today, and you'll be amazed how quickly you can create stunning dashboards on your own. I'll see you there.