BESO
October 25, 2021
Grupo Scorpion
November 4, 2021

http://www.grupovizion.com/

At a Glance

Challenge: Before switching to Google Cloud Platform, they had an on-prem server which only had the capacity to show averages for the last week, this being in addition to having slow processing and a high maintenance cost when trying to generate reports in order to deliver them to the systems team.

Solution: Migration process to Google Cloud Platform hand in hand with a Premier Partner, in this case, Nubosoft.

About grupo Vizion

Waldo’s Mart was a store of everything at $100MXN (Mexican Pesos) now it is a mini-supermarket where its differentiation strategy consists of offering several products, but no longer at a single price.. In 1999, it began operations in Mexico, starting in Tijuana, the first city that took the commercial coup from Waldo’s. The concept was so well taken in Mexico as well as in stores in the United States of America, that the decision was made to close stores in the United States and only work on expanding stores in Mexico. Thus, in June 2000, the central offices were established in the city of Tijuana.

Nubosoft

The evaluation, migration and support services for the project were provided by Nubosoft.

Nubosoft is a Google Cloud Premier Partner focused on the Mexican market and experienced in thousands of Google Cloud implementations.

Google Cloud Platform

Grupo Vizion is using BigQuery for the storage and analysis of data generated with Jupyter Notebook

 

The Challenge

Grupo Vizion’s business demands the constant surveillance and analysis of all their SKU’s (Stock-keeping unit) inventory rotation in all their stores, specially, for the ones with the better sales performance. Before switching to Google Cloud Platform, they had an on-prem server which only had the capacity to show averages for the last week, this being in addition to having slow processing and a high maintenance cost when trying to generate reports in order to deliver them to the systems team. Furthermore, not having the capacity to forecast data due to the lack of on-premises server capacity.

 

The Solution

Altogether, Sergio Ignacio Muñoz Iturralde’s (Lead Data Scientist) and Nubosoft Data Engineering team, decided to work with a total of 21 python notebooks in order to start slowly presenting progress reports until they could start organizing, analysing and forecasting the data. Later the legacy report (Microsoft Excel) was migrated to Jupyter to notebooks where the data can be better handled and its integrity could be assured by the systems team. In this new process, Bigquery data is exported to Jupyter notebooks where, after being analyzed, it is stored in the Google Cloud Store. At the same time a CVS load is generated in a Jupyter Notebook , it is loaded into Oracle, where around 20 notebooks work individually in order to generate forecasts more quickly together with their corresponding reports. They are now using Windows virtual machines in Compute Engine with the connection stability and bandwidth required for the design and analysis of the reports needed by their Business Analytics team.

 

The Result

Grupo Vizion’s team can now generate data forecasts in less time since tasks that would take hours now take minutes. The reporting of data, now that they are able to generate and store their CSV files using Cloud Cloud Platform’s tools, is done in a few minutes. Saving money thanks to efficiencies in their TCO (Total Cost of Ownership), improvements in their business processes and an improved data access thanks to the uses of BigQuery.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.