Global and Industry Leading E-Commerce Marketplace
Using customer and transactions data, two primary analyses were done to help increase marketing ROI
Customer Behaviour Analysis: This was done using a clustering algorithm to segment out purchasing behaviour based on recency, frequency and monetary value of each transaction. The segmentation allowed the client to identify unique strategies for each customer group. We also calculated customer life time value, retention rate, acquisition rate and churn rate per customer segment.
Geographical and Industry Segmentation: This was done using a multivariate regression technique. The analysis result outlined top industry and geographical areas that seem to have the highest correlation with revenue. Conversations with the client helped understand the possible reasons behind these relationships which can be taken advantage of in optimizing marketing spending.
*Example of a cluster graph
*Example of two regression equations
Real Estate Service Provider
Content and web analysis to maximize website traffic
Web Search Traffic Analysis: This was done by integrating a text mining analysis with Google keywords analysis tool. The results were used to determine which search terms are most used in relation to topics relevant to the business. Demographic data was also extrapolated to understand the type of individuals behind the search terms.
Content Analysis: By taking top competitors' web page content and running it through a text mining algorithm to determine the most frequently searched words. Additionally, an analysis of each competitor's website structure was done to see how to create the best in class website.
*Example of a demographic data
*Example of a word cloud
Major Automotive Manufacturer
Model design and cross-validation for implementation of third-party natural language processing platform
AAARL was engaged to provide advisory services to a major automotive manufacturer looking to implement a third-party tool to analyze customer feedback, written in free-form text. While the third party vendor provided a scalable and integrated solution to the natural language processing problem, it provided few to no resources in designing models, performing validation, and testing procedures.
AAARL assisted through prototyping models via machine learning tools in order to frame the performance and viability of the client’s efforts to incorporate natural language processing. AAARL performed significant testing processes through the development phase, identifying major processes and model changes that needed to occur for the client in order to achieve specific KPI’s. Furthermore, AAARL conducted extensive cross-validation work to ensure the consistency of results for the client as they proceeded into the implementation phase.
*Samples of simulated results of how accurate the model was at predicting certain variables.
Education Technology Company
Data Pipeline and Data Analysis Framework Audit
AAARL worked with the data analysis team to perform a comprehensive top-down audit of a data package from a company critical project. Performed model validation, data verification and cross-reference checks to identify and fix any errors along the data pipeline. Audited items include the data governance framework, data security, data analysis framework, data model, statistical model, codes, data outputs, data visualization and dashboards. This project was done under a tight deadline and the client was able to complete the required audit on schedule for their clients.
Major Asset Management Corporation
Pricing Optimization and Reporting
A major asset management identified an opportunity to create a dynamic pricing optimization system to price some 1000+ SKUs of inventory. AAARL built a three-stage optimization, prediction, and automation system to increase revenue by over 20% and gross profit by 45%. The project paved the way to various other analytics and business intelligence project within the organization.
AAARL fully revamped the reporting process in the organization, streamlining time for reporting writing as well as automating and consolidating reports for mid and upper management. Also performed various economic research and competitive analysis, including travelling to various cities/countries to understand best practices.
Ivey Business School. Western University.
Taught 18 hours worth of lectures to the Masters of Science in Analytics students on various technical topics in big data, database management, business intelligence tool and data visualization techniques. Graded Homeworks for each session.
Commercial Industrial Manufacturing and Distribution Business.
Using purchase transactions data, two main opportunities were analyzed to improve revenue targeting and pricing. Found specific product to raise prices out of 10,000+ skus
Customer and Product Segmentation Based Price Optimization
The goal of segmentation is to provide clear insights that drive strategic decisions related to each group in order to maximize their value to the business. RFM segementation were generated along with recommended actions to deal with each group.
Model Based Price Optimization
A linear transformation OLS model is generated for quantity as target variable using the price, products, the interaction between price and products as well as statistically significant control variables which take into account exogenous (external) effects.
Analyzing meals data, analysis was done to recommend efficiency gains in staffing through meal demand prediction, along with menu changes and operational efficiency gains.
Using the demand data from each meal, our team constructed a holistic view of the demand for each meal and each item on the menu during the days of the week. Using this insight to inform staffing for the nights, and menu changes.
Data Aggregation and Pipeline Building for Dashboards
Worked directly with the president to build executive dashboards on multiple store sales along with salary costs. Dashboard was built on Qlik Cloud and backend/datapipeline was developed from scratch.