Data-Talk-AI: “Chat with Your Data Files VIZ CSV, XLSX Effortlessly”

The task of collecting and examining raw data is often difficult for both people and businesses who do not have any technical knowledge on how to code or have the finances to purchase expensive data analysis tools. The current solutions require users to know how to use programming languages (i.e., Python and R) or require the use of cloud services where many users have concerns about privacy, cost, and access to the internet. Growing differences between the amount of data and the ability to use data led to the creation of Data-Talk-AI, which is a user-friendly, web-based application that will allow customers to interact with their datasets using everyday conversation in plain English while not writing any code. The Data-Talk-AI product uses a locally cached Qwen 2.5 (large language model)-built on Ollama- with a Streamlit UI, to provide an all-inclusive offline, privacy-preserving data cleaning and analysis solution. There are 11 steps Rule-Based automated data cleaning (missing rows/duplicate rows/outliers/sentinels/missing data type/mixed type/inconsistent format) with one button click. Users clean their data before requesting interactive visualizations via natural language, and the product automatically determines what type of chart to use/sets up axis labels & legend/configures the context based on the generated chart + provides the user with AI-generated context-based insights from the chart created. Data-Talk-AI uses MySQL-based authentication and bcrypt hashed password storage to operate fully on local infrastructure (no reliance on any cloud services) and provides non-tech professionals with professional-grade data management.