Problem
Over 70% of Malawians use mobile money (Airtel Money, TNM Mpamba) but have no way to track, analyze, or visualize their financial habits. Bank apps are inaccessible to most. The raw SMS transaction data sits unused on people's phones.
Solution
SmsSava reads exported SMS data (XML/CSV from Android backup apps), uses regex + NLP-based parsing to identify transaction types (send, receive, pay bill, buy airtime), amounts, recipients, and timestamps. It then builds a categorized spending profile and generates weekly/monthly reports as visual dashboards exported to PDF.
Real-World Impact
Tested with 3 real users. Identified average of 15% of monthly income going to untracked micro-transactions. Enables data-driven financial decisions for people outside formal banking.
Challenges Faced
SMS formats vary by carrier and change without notice. Building a robust parser that handles Airtel Money, TNM Mpamba, and NBS Bank formats required extensive pattern matching and fallback logic.
Key Learnings
Mobile money transaction data is rich and underutilized. Even simple NLP can extract meaningful financial intelligence from unstructured SMS text.
Demo & Execution Screenshots
