Spam Call Identification for Canar Telecommunication


Ebaa Marouf


Why Address Spam Calls?

Increasing spam calls received through Canar network

Importance of trust in telecom both with customers and partners telecom operators.

Reduce cost that Canar pays for partner network operator as fees to process calls in their network.

Tools Used

Clean Lined Data Analysis

SQL for data collection and partial processing

Microsoft Excel for data processing

Advanced Filter, Conditional Formatting, PivotTables, and PivotCharts

Methodology

Agile gradient icon
  • Data Collection: relevant data including historical call logs, inclusive of timestamps, origin numbers, and network ID has been collected from different systems.
  • Pre-processing: This involved removing anomalies and structuring the data for consistency.
  • Feature Engineering: up-normal behaviour like call frequency, timing deviations, and customer block requests to tag potential spam has been identified
  • Analysis: Data has been segmented to unveil patterns and to identify sources with high spam likelihood
  • Reporting: Final findings has been shared with stakeholders including evidences and recommended action points

Project Outcomes

  • Effective Flagging: the implemented solution flagged potential spam calls with a high accuracy rate, filtering out nuisances without impeding genuine calls.
  • Network-Level Action: Caller IDs identified as consistent spam sources were blocked at the network side. This decisive action ensured that no future traffic came from these malicious sources, significantly fortifying our defense against spam.
  • Complaint Reduction: Within the first quarter post-implementation, there was a notable 70% drop in spam call-related complaints.


Future Enhancements


  1. Automated Data Collection and Processing: Python can be used to automate data collection, ensuring real-time updates and faster data processing.
  2. Advanced Analytics: Using Python's rich ecosystem of libraries, such as Pandas for data manipulation and Scikit-learn for machine learning, will allow for more in-depth insights and predictive modeling.
  3. Integration with Other Systems: Python can serve as an intermediary tool, bridging Excel with other systems or databases, making the overall process more integrated and seamless.


The end