CASE STUDY 04
- Professional
- Generative AI
- BI
Querying data in your own words
An NL2SQL application for non-technical users.
Tressol-Chabrier · Work-study placement
The project, step by step
Natural languagecustomer master data exploration
- 01
Ask a question
In natural language
- 02
Generate the query
Mistral Large translates the question into SQL
- 03
Query the data
Access customer master data through pyodbc
- 04
Explore the answer
Plotly charts in Streamlit
Context
Development of a natural language analytics application at Groupe Tressol-Chabrier.
Problem
Enable non-technical users to explore customer master data without writing SQL queries.
Data
The group's customer master data. Its exact scope and volume are not specified.
Approach
I developed a Streamlit application combining natural language queries, SQL access and visualisation.
Method
The application uses the Mistral Large API, Python, pyodbc and Plotly to enable exploration of the master data.
Results
An NL2SQL application that enables business teams to query customer master data in natural language.
Technologies
Streamlit, Mistral Large API, Python, pyodbc and Plotly.
Lessons and limitations
Query reliability, handling of ambiguous questions and access control arrangements remain to be documented; no success rate is reported.