Data Preparation with AI Assistance: Quickly & Intuitively Merge and Transform Data

July 8, 2025

Video Tutorials
Merging multiple ERP tables into one file with the help of an AI chatbot.
Show Transcript

Hi! For preparing time series data, I use the CHECK-IN in future. The CHECK-IN requires that the data is in a single file. If your raw data is still in multiple tables, you’ll need to format the files correctly. In this video, I’ll show you how to easily and conveniently transform your data into the right format using two prepared tools. For this, we provide a prompt template and a Jupyter Notebook. I’ll show you the easiest way to use them right now.

I’ll start with an AI chatbot like ChatGPT, Gemini, or an internal company chatbot. For this video, I’ll be using Gemini. You don’t have to give the chatbot all of your data. Instead, it will ask you for the necessary information about your data’s structure, like file and column names, in a dialogue. Here, you should make sure that these names don’t contain any confidential information, unless you’re using a system that is specifically approved for it. This is what it looks like. You copy and paste the prepared template into the chatbot. Then it starts asking questions. First, it asks you for the names of your files. It’s best to just copy and paste the file path directly. I’ll call the first file ORDR for now. Then I’ll add the second file as well. The chatbot briefly confirms what it understood, and that’s the great advantage of this interview approach. If anything is unclear, the chatbot can ask for clarification, and you can respond. And vice versa, if you’re not sure what the chatbot wants to know, you can also ask.

I’ll switch to a chat where I’ve already gone through all the steps. So, after we’ve answered all the questions here, the chatbot creates a complete Python script. It also gives us some tips on how to use it and explains what it did in the background. You can simply copy this code block after you’ve reviewed it. This brings us to tool number 2, our specially created notebook. Here, you paste the code snippet, click the run symbol, and get immediate feedback. In our case, everything worked right away. The result is a single, prepared file. With that, we can start forecasting right away.

We’ll show you how to create the forecasts in another linked video. Or you can grab another notebook that best suits your case. For getting started, for example, the Getting Started notebook or one of the use case templates for sales or demand forecasts. One more thing for those who prefer to fill out the template with the information directly: You can find the template in the Notebooks folder, under Prompt-Templates. It’s basically a text template that explains the chatbot’s task. Instead of doing the interview, you could also manually fill this template with your information and send it directly to the chatbot. However, we find the interview approach more intuitive, and it has worked very well for us. That’s why we wanted to share it with you.

We look forward to your feedback. If you have any questions, you can always contact us at support@future-forecasting.de. Happy forecasting!

Quickly Prepare Your Data for CHECK-IN with AI Support

Here you’ll find the guide to our video tutorial, which shows you how to quickly and easily prepare data from multiple files or tables into the correct table format for CHECK-IN, without having to delve into the details of the required data format. An AI chatbot will help you create a Python script tailored to your data.

This preparation is the foundation for automated forecast generation with futureEXPERT and can also serve as a basis for a custom data connection.

Table of Contents

Your Starter Kit for Data Preparation: The Tools

For the transformation of your data, we provide two free tools that you can find directly in the futureEXPERT GitHub repo:

  • A Prompt Template: A template that explains the task to an AI chatbot (e.g., Github Copilot, ChatGPT, Gemini) and guides it to ask you the right questions.
  • A Jupyter Notebook for flexible data preparation: An interactive environment that guides you through the workflow & where you can directly execute the Python script generated by the chatbot.

The Data Preparation Workflow in Brief

The entire AI-assisted data preparation process can be summarized in three core steps:

1. Submit prompt to chatbot & answer questions ➡️ 2. Receive & review Python code ➡️ 3. Execute code in the notebook

Step 1: The Interview with the AI Chatbot (0:32)

Start a dialogue with a chatbot of your choice. The goal is to explain the structure of your data without having to share the actual data content.

  1.  Use the prompt template: Copy the content of the template (PromptTemplate.docx) and paste it into the chat.
  2.  Provide metadata: The chatbot will now start an interview. Answer its questions about the metadata of your files, such as file and column names. Your actual raw data remains private.

Tip: The interactive dialogue is the great strength of this approach. If something is unclear, the chatbot asks, and you can answer – and vice versa.

Step 2: Generate and Review Code (1:46)

After the chatbot has all the information, it creates a Python script tailored to your data requirements.

  1.  Let the code generate: Wait until the chatbot has created the complete Python code block.
  2.  Review the generated code: Treat the script as a draft. Check it for correctness, security, and whether it complies with your internal guidelines. This is a crucial step for which you are responsible.

Step 3: Execute Code and Get the Result (2:11)

Execute the reviewed script in the Jupyter Notebook to transform your files.

  1.  Insert code: Open the ai_assisted_data_transformation_for_checkin.ipynb notebook and paste the copied code into the designated cell.
  2.  Execute code: Click the “Run” icon (▶) of the cell.
  3.  Check the result: The code should now have generated a single, prepared .csv file in the correct format.

Next Steps: Creating Your Forecast

With your prepared file, you can get started right away:

Alternative for Experts: Fill Out the AI Template Directly (2:49)

If you want to skip the interview, you can also enter the information directly into the template.

  1.  Open the PromptTemplate.docx file in a text editor.
  2.  Manually enter the information about your files (paths, columns, etc.) in the designated places.
  3.  Copy the fully completed template to the chatbot to receive the code directly.

Further Information

You can find explanations for key terms from this workflow in our Forecasting Wiki:

Questions, Problems, or Ideas?

Your feedback helps us make future even better! Feel free to contact us anytime via email at support@future-forecasting.de or simply create an issue on GitHub.

Happy forecasting!

Last Updated: July 8, 2025
You are about to leave our website via an external link. Please note that the content of the linked page is beyond our control.

Cookies und andere (Dritt-)Dienste

Diese Website speichert Cookies auf Ihrem Computer nur, wenn Sie dem ausdrücklich zustimmen. Bei Zustimmung werden insbesondere auch Dritt-Dienste eingebunden, die zusätzliche Funktionalitäten, wie beispielsweise die Buchung von Terminen, bereitstellen. Diese Cookies und Dienste werden verwendet, um Informationen darüber zu sammeln, wie Sie mit unserer Website interagieren, und um Ihre Browser-Erfahrung zu verbessern und anzupassen. Zudem nutzen wir diese Informationen für Analysen und Messungen zu unseren Besuchern auf dieser Website und anderen Medien. Weitere Informationen zu den von uns verwendeten Cookies und Dritt-Diensten finden Sie in unseren Datenschutzbestimmungen.