Contract type: consulting
Client: Telefónica Germany Gmbh & Co
OHG
Role: Data Science Consultant
Time period: Feb 2017 – Dec
2018
Volume: ~ 2300h
data science • Kyoto Cabinet • Linux • docker • NoSQL • Python • matplotlib • numpy • scikit-learn • scipy • QGIS • apache hadoop • apache hdfs • apache hive • apache hbase • apache phoenix • apache kafka • tableau
I developed a tool to identify, through text mining, tickets in a ticketing system at Telefónica which need to be brought to the attention of management.
This project was based on data collected by Telefónica Germany, Germany’s largest mobile network operator, within their network operations ticketing system.
Using techniques from statistical natural language processing, I created a model to predict, based on the text content of a ticket, whether or not it is likely to take a long time to resolve.
The final deliverable was an alerting system that was able to automatically send e-mails to managers when business processes fail to run smoothly or otherwise need attention.