Tools: Looker, dbt, snowflake, fivetranDomain: Ecommerce and Logistics Analytics

At Infrastructure NSW, I led the setup of the analytics function to improve evidence-based risk assessment for large-scale infrastructure projects. I managed multiple concurrent projects and drove initiatives in a cross-functional environment, coached and mentored team members and external consultants, and reduced time from insight to action by developing a Command Centre as a one-stop-shop for all key metrics and insights

Hassam delivered projects involving anomaly detection, demand forecasting and propensity modelling.Anomaly Detection: - Used a combination of supervised and unsupervised techniques to identify outliers within stores and agents that may have a higher chance of committing fraud.Demand Forecasting:Business Problem: Establish sales targets for stores based on their historical performance and future market potential - Used trend forecasting and causal forecasting techniques to optimize sales targets generation for brick and mortar stores at a telecommunication company.Propensity Modelling: - Designed a lead scoring model for an education services provider to help them qualify high volume leads coming from emerging markets. Also deployed a “likelihood to confirm” model that helped their direct sales team to use the right levers to push a student from offer to confirmation stage.

Advance Segmentation: - Used unsupervised learning techniques (k-means, hierarchical) on email and transaction behavior to identify key customer personas for retail clients.Market Basket Analysis: - Analyzed customer affinities for different product categories to identify cross-sell/up-sell opportunities. Used association mining techniques such as apriori on transaction data and website visits to identify products that were bought frequently together..- Optimised the conversion rates for key customer journeys. Used linear regression to identify top factors influencing the conversion rate.Response Modeling: - Used algorithms such as random forest/xgboost to identify customer propensities towards certain offers. This enabled the business to improve the conversion rates for their offers.

Key achievements:- Used linear regression on qualitative survey data to identify factors that affect overall satisfaction rates for Audi Australia customers. Completed a driver analysis for NPS. Highlighted attributes such as cost of service, vehicle satisfaction, ease of booking, appearance of service area etc. and their effects on customer satisfaction and loyalty.- Used CHAID (Chi Square Automatic Interaction Detector) to identify variables such as enrollment date, number of emails opened in last 60 days, days since last email opened, etc. that can predict the likelihood of an Audi customer unsubscribing from email marketing.- Optimisation of all Audi Australia websites and microsites using data from Google Analytics, Omniture and Kissmetrics. Carried out A/B testing of key goal conversion funnels such as brochure downloads and test drive requests.- Ongoing provision of insights into the online behavior of Audi owners, leads and prospects and driving UX design modifications using data. Usage of statistical tests such as t – tests and two proportions test to verify whether any uplift in conversions is statistically significant.- Customer segmentation using Salesforce data to identify audiences for major sponsorship campaigns.- Behavioral analysis of Audi owners using qualitative data from Roy Morgan and ARMs.- Post campaign analysis of Audi model launch campaigns, events, sponsorships, retail offers, etc. using Google Analytics, Omniture, Salesforce and Marketo.- Ongoing development of Tableau dashboards to report key business metrics using Sales data, survey data, customer demographics and psychographics data from ARMs.- Used linear regression to identify drivers that affect satisfaction rates and customer loyalty (NPS).

- Used donor attributes such as gender, employment status, monthly gift amount, etc. to create a donor risk profile and predict whether they would cancel their monthly donation subscription after 100 days with a non-profit agency. Used logistic regression to develop a model for classifying a donor as high or low risk.- Used five years of student data, to identify variables that can predict the likelihood of a student dropping out of school. Used data mining techniques such as logistic regression, decision trees to identify factors such as effort scores, absence rates, sibling at school and financial situation that can predict student attrition.- Developed a platform for reporting and data visualization using Highcharts.js, PHP (Yii Framework) and MySQL.- Automated data extraction and data sets creation using Python scripts and SPLUNK.

- Conducted an audit of Australian Award websites.- Identified usability issues and outlined a strategy in line with the communications strategy.- Developed an analytics strategy for measuring key performance indicators for the websites using Google Analytics.

Worked on a real estate platform to buy, sell and advertise house and land all over Australia. Enhanced the search functionality to allow users to search for different real estate properties. Core work done in PHP, Zend Framework and MySQL.

Key Responsibilities:- Identified risk and weaknesses in existing I.T infrastructure of a client enterprise.- Designed and suggested improvements in the business processes of a large scale enterprise as a part of an overall Business Process Management strategy.- Reduced the post-rollout registration time for employers by 50% and for employees by 95% in the EOBI Facilitation System Project.- Worked on architecture and framework design of existing application infrastructure of a client.- Worked on a product catalogue application and a recommendation engine that allows Intel, Yahoo, Bestbuy to sell their laptops, desktops and netbooks to their consumers.- Worked on IBM Websphere Suite, a set of enterprise tools used for integration and business process management purposes.