Skip to content
Vijay Work Resume Blog Contact

Project case study

Retail category uplift and cannibalization

Category performance analytics for measuring launch uplift, incremental sales, and cannibalization effects.

Helped stakeholders understand how launches changed category performance beyond top-line sales alone.

Spark Python Power BI Airflow Apriori

Context

The problem

Retail stakeholders needed a clearer way to understand whether a new product launch grew the category, shifted demand from nearby products, or did both.

Combined statistical analysis with reliable data preparation so the output could be trusted and reused in business planning.

System trace

How the work moved through the system

A high-level operating path: where the request starts, how the system shapes it, and how other teams consume the result.

  1. 1

    Spark and Python prepared historical sales, product, and category-level datasets for measurement.

  2. 2

    Statistical analysis compared launch behavior against relevant category and product baselines.

  3. 3

    Power BI provided a consumable view for reviewing uplift and cannibalization patterns.

Business question

Uplift vs. shift

Separated incremental category contribution from demand that moved between related products.

Output style

Planning-ready

Prepared results for business planning conversations rather than only technical validation.

Architecture

System shape

3
  1. 1 Spark and Python prepared historical sales, product, and category-level datasets for measurement.
  2. 2 Statistical analysis compared launch behavior against relevant category and product baselines.
  3. 3 Power BI provided a consumable view for reviewing uplift and cannibalization patterns.

Ownership

What I handled

3
  1. 1 Built reliable data preparation for product launch and category performance analysis.
  2. 2 Supported measurement logic that distinguished incremental contribution from cannibalization.
  3. 3 Packaged outputs so retail stakeholders could evaluate launch outcomes and tradeoffs.

Lessons

What carried forward

2
  1. 1 Launch analytics need a clear baseline story or the output becomes easy to misread.
  2. 2 Cannibalization is a data-preparation problem as much as a statistical one because product relationships define the question.

Engineering decisions

Measure the category, not only the launched product

Looking only at the launched product could overstate success, so the analysis considered adjacent products and total category behavior.

Keep the workflow reusable

The design favored repeatable preparation and reporting so similar launch evaluations could be run again with less manual work.

What can be shown

Public evidence without internal names

The internal systems stay private. This section keeps the public parts: my role, system boundaries, technology context, scale, decisions, constraints, and what I learned.

Internal enterprise system High-level architecture Scale signal

Business question

Uplift vs. shift

This case study can describe the measurement distinction without exposing launch or client data.

Architecture shape

  • Historical sales and product data is prepared through Spark and Python processing.
  • Measurement logic compares launch behavior against category and adjacent-product baselines.
  • Power BI provides planning-ready review of uplift and cannibalization patterns.

Responsibilities

  • Built data preparation for launch and category-performance analysis.
  • Supported measurement logic for incremental contribution and cannibalization.
  • Packaged outputs for business planning while keeping proprietary data private.

Constraints

  • Client names, launch calendars, product identifiers, baselines, and measurement rules are confidential.
  • Site notes use generalized category-performance language.

Supporting context

High-level architecture

High-level category measurement workflow

Can be shown as historical preparation, baseline comparison, uplift/cannibalization measurement, and BI review.

Scale signal

Measurement pattern signal

The case study can state uplift versus demand-shift analysis without publishing proprietary values.

Related case studies

Continue through related work or return to the full project index.

All projects

Related projects

Continue in the same area

Project index

Spark + Python + Data platform

Retail association-rule engine

Built a repeatable Apriori-based workflow for analyzing how products, categories, and departments are shopped together at scale.

Spark + Python + Data platform

Retail adjacency and store-flow analytics

Built reusable analytics workflows for cross-shopping, category adjacency, aisle-flow, and store-flow analysis across departments, categories, and products.