Azaz Rasool

Operationalization of P&L for AI & Data Analytics department.
AI & Data Service catalogue
P&L Structure
Financial model (Costing, Pricing and Payment models)
Pilots with internal Business Units and Subsidiaries

#1 Designed Strategy and Roadmap for AI & Data for 3 years (2021-23)
#2 Alignment with Business stakeholders on Business impact, ROI, CAPEX and OPEX
#3 Use case identification, prioritization and detailing
#4 Business case building
#5 Strategy execution, Target operating model, KPI definition and monitoring

Architecture Design and implementation of Enterprise Data Provisioning Platform on Cloudera Hadoop
Data Ingestion design patterns for Mainframe COBOL data files, CSV and RDBMS tables to Hadoop
Data Extraction design patterns for Consumer applications

Build taxonomy and dictionary of keywords and sentiments
Load chat files in Hadoop HDFS
Process chat files, extract sensitive words expressing customer sentiments
Load extracted keywords and bucket then as positive , negative and neutral sentiments , in separate tables/files
Visualization/reports in Tableau

PPV is an important business line for any cable TV provider and PPV ads are placed in view of increasing the sale of PPV.
These ads are aired in various part and time of the day. Since Ads get limited slots, if not placed rightly then will not result in conversion and hence loosing the revenue.
There was a need to Predict ad performance, Minimize ad spend, Utilize ad slots efficiently, Maximize PPV purchases and Understand Ad effectiveness by channel
Factors considered: Actual ad… عرض المزيد

PPV is an important business line for any cable TV provider and PPV ads are placed in view of increasing the sale of PPV.
These ads are aired in various part and time of the day. Since Ads get limited slots, if not placed rightly then will not result in conversion and hence loosing the revenue.
There was a need to Predict ad performance, Minimize ad spend, Utilize ad slots efficiently, Maximize PPV purchases and Understand Ad effectiveness by channel
Factors considered: Actual ad air time, Channel ad aired on, Window break, Break position
Models used: nPath, Logistic Regression, Decision Tree عرض أقل

* High customer churn (1.8% per month) was a cause of concern for the business which needs to be analyzed and addressed.
* Health of the set top box (STB) was believed to be an important factor for customer churn.
* There was a need to predict the churn behavior using STB data and accordingly proactive measures to retain the customers can be taken.
Models to separate churners and non-churners, Correctly identified 77% of churners and 97% of non-churners
Making Churn prediction… عرض المزيد

* High customer churn (1.8% per month) was a cause of concern for the business which needs to be analyzed and addressed.
* Health of the set top box (STB) was believed to be an important factor for customer churn.
* There was a need to predict the churn behavior using STB data and accordingly proactive measures to retain the customers can be taken.
Models to separate churners and non-churners, Correctly identified 77% of churners and 97% of non-churners
Making Churn prediction actionable: Roll service truck, Proactively contact customers at Risk, design response scripts
Factors considered: Days since last health report, Longevity of account, Services subscribed to, # of reboots, Occurrences of low/mid/high signal strength
Models used: Logistic Regression, Single Factor analysis, Decision Tree عرض أقل