Claude has introduced a new built-in feature that enables end-users to write and execute JavaScript code. While it previously had the code writing feature, reading and running codes using complex math calculation procedures and data analysis were missing from Antropic’s offering. 

The new data analysis is like a built-in code sandbox. It provides an environment where you can analyze data and iterate on different ideas before getting an answer. Now, you can get accurate answers when processing information and running codes. All this is based on the Claude 3.5 Sonnet’s data and coding skills. 

Whether analyzing and visualizing data from CSV files or executing JavaScript codes in real time, Claude produces precise outputs. It works like a real data analyst, processing your data systematically—cleaning, exploring and executing it. Claude analyzes the code step-by-step, providing detailed insights into its execution. 

Even though Claude can write code, it can not run code within its system. It supports all analytical tasks using its AI brain to give mathematically precise, reproducible, well-rounded answers.

We have devised a couple of scenarios to give you an idea of how to use the data analysis tool. We will show you how the tool helps marketing, sales, finance, and engineering teams analyze and visualize data. 

Key Objectives

1. Visualizing Marketing Data 
2. Analyzing Sales Data
3. Analyzing Finance Data
4. Visualizing Engineering Data

Visualizing Marketing Data 

Go to Claude.ai and click the account icon at the bottom left corner of the screen. Click your account email to access the menu. Select ‘Feature Preview’

You have to turn on the analysis tool. Toggle the button to turn it on. 

Close the pop-up window and visualize the marketing data using the following prompt. We have a sample marketing file in JavaScript. At the time of writing this tutorial, Claude supports the analysis of JavaScript or CSV files only. 

Prompt:

Can you please visualize the data and separate it so I can find bottlenecks in the marketing campaign? 

Note: Attach a Javascript file. It shouldn’t be a large file. The preview analysis tool doesn’t support large files. 

‍

Claude will extract the code from the file and use it to visualize the data based on the given prompt. Here is what it generated from a sample marketing data file. It identified the bottlenecks and provided recommendations.

You can view the code by clicking the code button at the top of the visualization pane. 

You can copy, download or publish the data by clicking the button at the bottom of the visualization pane. 

You can always follow up with additional prompts to get the most out of the data. For example, we asked Claude to visualize the bottlenecks, too. Here is the prompt:

Prompt:

Can you please provide visualizations of the bottlenecks you identified in the data?

You can see that it provided bounce rates for each channel visualization that are integrated with the current visualization. 

Analyzing Sales Data

Visualizing data using the Claude analysis tool was fantastic. It can also analyze various data types, including data in CSV files. Upload a file and write a detailed prompt asking Claude to analyze various data points. Here is a prompt for analyzing sales data.

Prompt:

Can you please analyze sales data and determine the number of sales we had in a month? Also, provide the grand total of the money collected and other data points.

It read the contents of the file in JSON, cleaned and parsed the data and provided the required analysis with insights.

It showed the total number of transactions and money collected, which was very helpful. The additional insights were satisfactory but not very beneficial.

Of course, you can always follow up with additional prompts to get meaningful insights from your data. For example, you can find more data points by asking Claude to analyze:

• The modifier patterns to see what add-ons are most popular
• Create a visualization of your sales timing to help identify busy periods
• Analyze the size distribution between regular and XL orders

It depends on what you want. Do not hesitate to ask follow-up questions!

Analyzing Finance Data

Finance people spend countless hours analyzing financial data, which can sometimes take days to find useful insights. With Claude's data analysis tool, analyzing or visualizing financial data can be done in a minute. 

We have some finance data in a CSV file. Let’s see what Claude can do with it. Use the following prompt:

Prompt:

Please analyze the data in the file and provide a detailed analysis of the expenses and liabilities and how they impacted the company's profits.

It did a nice job analyzing the gibberish in the financials file. It analyzed the key points and provided a deep analysis of the data. 

You can create a financial dashboard using a data analysis tool. The dashboard will look like this:

Again, it depends on what you want to do with the data. Not only can it create data dashboards, but it can also generate key metrics, interactive account-type transaction trends and other data points. You can instruct Claude to add it to the data dashboard. 

Visualizing Engineering Data

Engineering data is complex and requires hours of analysis to find meaningful insights. However, the Claude data analysis tool allows you to perform analysis and data visualization simultaneously. 

Our scenario uses a JSON file filled with copper wire manufacturing data. Here is the prompt:

Prompt:

Can you please provide detailed insights into the number of personnel needed and the total run time for each machine for the entire available material?

It created a fantastic visualization in a bar graph and provided the insights we wanted from the data. 

To create a data dashboard of the insights it provided, use the following prompt:

Prompt:

Please create a dashboard of the data and the resources used to produce the copper wires from the raw materials. 

The dashboard is great. We loved how Claude arranged the data insights and generated a pie chart and a bar graph of the entire dataset. 

There you have it. Do not hesitate to experiment with the data analysis tool. It might be a bit limited in supporting various file types, but the output is impressive, to say the least.