ChartPixel is an AI-assisted data analysis platform that helps users create visually appealing and informative charts and insights from raw data in seconds. It uses a variety of AI techniques, including natural language processing, machine learning, and data visualization, to automatically clean the data, select the most suitable chart types and columns, and generate charts and insights that are ranked by the importance of the insights found in the data.

ChartPixel is easy to use, even for users with no prior experience in data analysis. Simply upload your data file and ChartPixel automatically generates the most important charts and insights for you. You can then customize the charts to your liking, including changing the chart type, colors, and fonts.


ChartPixel has been a game changer for students, teachers, researchers, analysts and business professionals at all skill levels.

Chris: Hi there, please introduce yourself.

Jack:  Hello, I’m Jack Witkowski, and together with my co-founder Andrea Szilagyi, I launched ChartPixel – an AI-assisted data analysis platform. Our tool automatically generates charts and written insights from raw data.

I hold a Master’s degree in Econometrics from Oxford University. After completing my education, I ventured into various fields but always maintained a strong affinity for data. Initially, I worked as a lecturer, teaching Data Analysis to students. Later I continued my career as a hands-on Data Scientist and Data Storyteller in London and Barcelona. Each of these experiences inspired me to create my own startup, and in hindsight, they have all significantly contributed to what ChartPixel is today.

Chris: How did you come to create ChartPixel?

Jack: It was classic. I was looking for a remote job in Barcelona. As part of the interviewing process, spreadsheets are often emailed to the applicants, who are tasked to analyze the data and present the findings to the interviewing board. It is nothing too complex, but this process can be quite tedious as the files can vastly differ from each other; from questionnaires, over marketing to financial data. After a few weeks of applying to various jobs, I was sitting on a mountain of data files.

Before I could even create basic charts, the data had to be cleaned and reshaped. Simple correlations needed to be checked, and basic insights, such as statistical trends and patterns, had to be extracted. Although each file was unique, the process felt repetitive. I began to look for a tool that could streamline the results – something that could fastrack from uploading raw data to generating comprehensive charts with accompanying explanations suitable for sleek presentations.

This is where I noticed a gap in the market for quick and smart data exploration. I scripted my first line of code which grew over time and at some point grew to a more cohesive codebase. When I uploaded a large questionnaire and the code generated over 100 charts, ranked from good to bad, and with basic descriptions, I knew I was onto something.

Chris: Please tell us about your experience in the AI field.

Jack: I did a 10-week course in 2020 at Oxford on how to build AI applications. The course went into great depth explaining the maths of deep neural networks, LLMs, and how applications in the AI field have evolved over time. This gave me a good headstart on understanding the architecture and needs of AI when building an application.

Chris: What algorithms and techniques does ChartPixel use to generate charts and insights?

Jack: ChartPixel uses a blend of AI and rule-based domain knowledge. It was important for us not to solely rely on AI, but to have AI assist us in leading the analysis in the right direction and adding context.

For example, to detect if a user has uploaded a questionnaire would be far too complex to rely on a rule-based approach. This is where the strength of deep neural networks and LLMs comes in handy since column headers and types are better understood in their entirety. The same applies to generating color themes, data-driven insights, and speculations.

We provide the framework where AI can safely play with rules that we have carefully crafted and predefined.

Chris: How does ChartPixel ensure that its results are trustworthy?

Jack: We ensure that domain knowledge and statistics go hand in hand with AI. However, when in doubt, domain knowledge trumps AI. While our AI or ML algorithms may suggest appropriate statistical tests or filter out charts deemed to be of low quality, we generally avoid allowing the AI to create new tests or generate additional charts. Essentially, we utilize AI to narrow down existing options rather than expanding them. It is a funnel in our opinion and not a broadcast.

Chris: What makes ChartPixel unique from other data visualization and analytics tools?

Jack: ChartPixel proactively provides written insights and descriptions that are suitable for skill levels, without requiring you to ask questions. The platform educates you on not just what the chart shows, but also its implications for your business or research. Our focus extends beyond merely generating charts; we aim to produce actionable insights underpinned by statistical analysis.

The platform automatically selects pertinent columns, tidies up messy data, and even engineers new features to streamline your data analysis process. Within seconds, you can effortlessly export your analysis to PowerPoint presentations or share the comprehensive analysis with your peers. Designed for intuitiveness, ChartPixel eliminates the steep learning curve often associated with other data analysis tools.

Chris: How do you think ChartPixel is changing the way people interact with data?

Jack: We want to make data analysis both available and fun for anybody that holds data. We find that data becomes more alive when there is an inherent smoothness in the data analysis process.

Traditional data analysis and visualization tools are flooded with an overload of options, customizations, tests, chart types to choose from, and so on. These tools are driven by the following principle: The user starts with a goal or end result in mind. They learn what options are available to them by slowly familiarizing themselves with the tool. Then they work their way up to reach their goal by selecting and filtering through all these options.

With AI and its intuitive capabilities, we are experiencing an interesting paradigm shift. For the first time, we can actually present the end result to the user’s goal right at the beginning of their journey. AI has made all the selections for you based one some initial parameters. The user can then modify or customize elements that are the building blocks that lead to the final result. This AI-driven approach is inherently different from how traditional tools work.

One can think of it like this: You want to create a computer game, but you are not a programmer, or, you got inspired today about what happened and want to write a novel but you are not a good writer. Unless you want to learn a lot, this would normally be the end of the road. But now we have AI. You can outline the plot that inspired you and then set some parameters like location, time, and main characters. The AI can write the entire novel for you. Then you read each chapter and specify more parameters, delete some characters, or add new story elements and you can do this until you are happy.

Why is that fun and more importantly why is this working out well? Here comes the interesting bit: Only a few people have the skills to write a novel or code a game. But everyone can read a book and judge if it has been inspiring, or play a game and understand if it is a good game and is fun to play. So, even if you are not a writer or a programmer, you have now the opportunity to accomplish great things with the help of AI blended with your judgment and human input.

CharPixel aims to do this with data analysis and empower users. You upload your data and you are already presented with your end result – an analysis of your data that is suitable for anyone to understand. Then you look at some specific charts and want to change the format or colors, dive deeper into the insights or dismiss and when you are happy you convert it into a sleek presentation.

The end result of data analysis is to extract insights that are presentable to and understandable by anyone, for example, the board of directors. The board might not know statistics or how the analyst got there step by step, but they understand the charts and insights and how to use them in their business. Moreover, they can trust the results and know there is some sort of check or framework in place.

Chris: What are some of the challenges you’ve faced in developing and launching ChartPixel?

Jack: As a self-funded startup, we took on the challenge to compete with VC-backed analytics tools and the big players in the market including some of the big ones such as Microsoft Excel, Power BI, ChatGPT, Tableau, and SPSS.

Chris: Thanks for being with me, any last words? Where can our readers follow you?