Many micro-enterprises go into crisis without any obvious signs until it's too late. This project aimed to create an intelligent warning system capable of detecting early signs of economic and financial risk, enabling timely intervention by business associations and managers themselves.
- Goal
Anticipate situations of financial and operational stress in micro and small companies, based on simple but critical data - such as sales, stock variation, customer turnover and compliance with tax obligations.
- Technologies and Tools
- Python
- Machine Learning (Decision Trees, Logistic Regression)
- Pandas
- Excel
- Power BI
- Innovation and Impact
- Imminent cash shortage
- Progressive loss of customers
- Systematic tax noncompliance
- Sustained drop in turnover or margin
The system combines business rules with predictive algorithms to generate automatic alerts and practical recommendations. The association thus takes on a proactive role in supporting its members - acting before the crisis hits.
- Results and Impact
In a pilot with anonymized data from 30 companies, the system achieved an accuracy rate of 78% in predicting financial stress 2 months in advance. The companies supported were able to implement corrective measures, with direct gains in liquidity and internal reorganization.
Português 
English