Retail analytics framework

  • Country


  • Industry

    Data Science

  • Type


  • Duration

    6 Month


A fast-growing apparel company with a focus on promotional and seasonal sales needed more advanced analytics to better support their retail model. Whatever the nature of a sales campaign, they wanted more precise customer information—such as buying behavior and likely preferences—and be prepared to offer targeted items, convert on the spot, and keep their customers coming back. For this, our client needed an additional component: a more advanced recommendation engine.


As a first step, our client’s data—which comprised customer purchasing history and behavior, and included unstructured text descriptions, comments, and reviews—needed work. We ingested the data and warehoused it to produce a coherent and modeled data warehouse. We then developed a recommendation engine to help predict what to offer their customers, the pricing, discounts, and so on—based on the aggregated data and the analytics derived from it. We also addressed the “cold start” problem by combining the content-based recommendation approach with next-purchase predictions based on customer profiles. Additionally, we developed a set of tools to help predict demand and gauge the effectiveness of their marketing campaigns.

Technical details

  • R
  • Python
  • NLTK
  • Scikit-learn
  • Java
  • Spring Boot


  • improved analytical insights the marketing staff could apply to devising targeted promotions;
  • increased by 5% average volume of order (AVO) and repeat customer rate (RCR).

    Contact us

    Interested in working together? Ok, don't be shy. Just fill out the form below.

    Don't like forms? That's ok, just email us.