Industry Challenge/ Challenges
With its high-growth, unique hypermarket strategy, the client wanted to know how they might use the ten-year historical data they had accumulated. Algoscale left no stone unturned in gathering and delivering relevant data analytics, allowing the client to make strategic decisions such as determining the optimal location for opening new stores to fuel the company’s growth.
- Poor data warehousing
Data warehousing at the client-side was poorly maintained. - Proximity
Store proximity was another challenge that needed attention. Geospatial data of the local population was recorded by the company. - Positioning
Locations of stores needed to be strategically placed.
Problem Statement
- To leverage data analytics and enable this hypermarket chain to predict sales for pipeline stores across Russia using geospatial data.
Our Solution
- We began by retrieving information from the client database and the AWS cloud. While recording data analytics, a protocol was implemented that involved the introduction of a new data nomenclature system as well as constant data quality assurance.
- Geospatial analysis was performed pre-processing multiple data sources, the results of which were drawn out of the location-mapped population-catchment data. Competitor’s store locations were also considered to avoid cannibalism.
- For predicting sales and footfall of pipeline stores, a Hybrid Prediction Model was run and population-sales behavior reports were created to acquire insights into store preferences of diverse economic groups.
Tech Stack
- Python and R
- PostgreSQL
- AWS S3 cloud for data storage
- AWS EC2 for Server creation
Business Impact
- The heatmaps of the catchment area provided the relevant details about areas where customers were declining and where new customers could be found.
- Our solution provided unprecedented insight into store performance, resulting in a drop in inter-store cannibalism and an increase in new cardholders for new stores.
- Based on population-mapped catchment and cannibal data, the client was aided with the ability to locate new stores and the geospatial analysis enabled sales forecast for pipeline stores across Russia.
