Success Story Telecom

Telecom Network Provider Reduces Data Quality Issues by 35%

Develop an end-to-end, full-scope EDW-BI application

An American telecommunications company that provides wireless services, is an internet service provider, and is the fourth-largest mobile network operator in the United States wanted to monitor the performance of the telecom network in near real time.

Our solution was to develop an end-to-end, full-scope EDW-BI application that included data modeling, ETL, Cube, and reporting framework, with a centralized data store as a single source of truth, resulting in improved operational performance and faster data processing.

the Results

Key Outcomes

At the close of the project, Apexon was able to deliver the following features and upgrades:

Improved service-network performance CSAT by 14.29%
Improved service-network performance CSAT by 14.29%

Resulted in 35% reduction in data quality issues
Resulted in 35% reduction in data quality issues

20% increase in governance & regulatory compliances
20% increase in governance & regulatory compliances

Due to real-time, network-cluster analysis
Due to real-time, network-cluster analysis, service downtime was reduced drastically

Improvised business analytics
Due to improvised business analytics, business rules could be modified & implemented seamlessly, improving gross profit by 5%

New system handles 1.9 PB of data enabling seamless analytical processing
The new system handles 1.9 PB of data enabling seamless analytical processing on it

New centralized datastore enabled
New centralized datastore enabled business to easily get 56+ report/dashboards

The challenge

4 key areas

Aside from the challenges associated with BI functional areas as data was scattered across different LOBs, Apexon specialists discovered that:

In-memory Processing

In-memory Processing PLSQL modules were used to design data aggregation in certain LOBs. With the increase in data volume, the business was concerned about job execution time and frequent failures caused by in-memory processing

Demographic Services Data

Demographic Services Data New demographic services data was required to be collected and maintained in a resource-storage and cost-effective manner to cater to certain analytics requirements

Switch-Network Logs

Switch-Network Logs There was no collection of Switch-Network logs (provided by vendors such as Nortel, Motorola, Samsung, and Lucent) and no scope for real-time network-cluster performance

Centralized data store

Centralized data store The client had no centralized data store to house customers service tenure details

The Solution

5 key areas

We implemented the following solution to meet the customer’s requirement of monitoring performance in near real time:

Cubes Designed & Built

Cubes Designed & Built

Various Cubes were designed and built with appropriate partitioning, aggregation, and caching strategies to perform capacity, performance, and network stats trending

Splunk Built & Maintained

Splunk Built & Maintained

Built Splunk dashboards and maintained certain customer and demographic services data into SNOWFLAKE cloud data warehouse, building snowsql reports and chartio dashboard

Data Cycles

Data Cycles

The data cycle extracts data from binary files in mediation servers and checks for data gaps and quality in the multi-threaded mediation and ETL stages

End-to-End ETL

End-to-End ETL

Built an end-to-end ETL solution that ingested data into Hadoop as well as landing TD tables. The data is also being ingested into Splunk

Full-Stack EDW-BI Application

Full-Stack EDW-BI Application

End-to-end design of a full-stack EDW-BI application, including data modelling, ETL, cubes, and reporting framework