IT Ticket Monitoring & Prediction IT Ticket Monitoring & Prediction

potential3

Predictive Analytics

The question: How many customer incidents will be received within the next two weeks?

For this question, INFONEA Data Scientists concentrate on deriving predictions for the future evolution of incoming incidents – simulated by workload and simulated by staff. The data basis consists of IT tickets i.e. customer incidents and notification of problems. These tickets are dealt with by customer service:

  • The customer service consists of various departments dealing each with a certain category of tickets.
  • Each department counts only a certain number of employees.
  • Each ticket requires a certain processing time.
  • The IT staff has fixed working hours from Monday to Friday.
  • Each employee may drop out from time to time (vacation, illness) which leads to fluctuations within the number of available employees.
  • Tickets with a higher priority are processed first.

This leads to numerous attributes for each ticket such as:

  • Category
  • Priority
  • Department processing the ticket
  • Check-in time of the ticket
  • Start of processing the ticket
  • End of processing the ticket
This data is then analyzed in INFONEA. Next to the common dashboard, INFONEA enables a look at the history and the analysis for patterns and anomalies within the ticket data.
 
Screen of INFONEA dashboard
INFONEA Predictive: Dashboard
Screen of  Patterns and Anomalies
INFONEA Predictive: Patterns and Anomalies

In the context of "Predictive Analytics", the INFONEA Data Scientists had the goal to make predictions regarding e.g. the waiting period of tickets or the evolution of the “throughput quota” (number of tickets that can be processed on the same day). In addition, hypothetical questions (“what if”) should be answered:

  • How does the average waiting period of tickets change in the next two weeks, if 50% more tickets were coming in within that time?
  • How does the throughput rate change, if 20% more/less employees were available?



potential3 dotted line
Predictive Analytics Screen