Case Study
The Client
The client is a leading provider of telecommunications and mobile money services, with a presence in 14 countries in Africa, primarily in East Africa and Central and West Africa. They offer an integrated suite of telecommunications solutions to its subscribers, including mobile voice and data services as well as mobile money services both nationally and internationally. The group aims to continue providing a simple and intuitive customer experience through streamlined customer journeys.
Project Objective – Automated Data Translation Solution
Business Values – Niveus is helping the client by providing an automated data translation solution for translating bulk data, thereby doing away with cumbersome manual tasks, saving time and streamlining effort.
The client was looking to translate their data related to their invoices, assets, inventory, and journal vouchers that are daily generated in bulk. The client has multiple entities in different countries, most notably in French-speaking countries/regions and so they required an efficient data translation solution.
Business Solution
Niveus has helped the client to translate approximately 10,000 rows of data in English to French. This data is translated from a single column which contains the text string description leaving other columns untouched. The solution includes scope for adding more columns and details for translating more data in the future.
Implementation
- Cloud Scheduler schedules the daily ingestion of csv export file from the client’s on-premise server
- The source csv would be stored in Cloud Storage bucket
- Cloud Function converts csv to supported file format for Translation API
- Google Cloud Dataflow calls on Cloud Translation API to translate from source language to target language
- The translated file was stored in a separate bucket (bucket 2)
- BigQuery was used to store csv files in the form of a table. It could also be explored to build AI/ML use cases for future by storing the translations completed
- The translated csv is pushed to the client database