careerbuilder-original-2992

Predictive analysis & portable platform integrating added-value services for restaurateurs

About the Client

Headquartered in Germany, the company works to provide innovative data-driven services for restaurateur and hospitality industries. To-date they have supported the online growth of over 220,000 restaurateurs in over 16 countries. With competitive solutions and products meant to aid in online profit growth (such as online ordering and reservation systems) they allow allows their partners to focus on what matters most: individualized customer service for their patrons.

 

The Goal

Create a unified customer view to facilitate (predictive) analysis and a flexible, scalable and portable platform supporting added-value services for restaurateurs. 

 

2-FoodServices_Transforming_with_Tech

The Challenge

Together with a partner company, Crystalloids was hired to address a number of challenges. Most pressing was that with their current data warehouse our clients' data scientists were unable to perform proper analysis as none of the data was centralized. Our data engineers had to find a way to connect all data sources in a proper structure so that the client no longer needed to create their own logic to connect data sources together. Unable to deliver a 360 consumer profile with their current data warehouse structure, our technical expertise in data engineering proved critical. 

Throughout our four-year partnership with this this retailer, our team was tasked with engineering a data warehouse that could connect 25 sources of data. Our team had to also create various cloud functions to provide data to other departments allowing them the capability to fetch data from the warehouse directly. Our team also had to integrate their current tools to BigQuery so that the data could be pulled for analysis. We had to work with systems outside of Google Cloud, such as Salesforce, and then extract all the data using Google Cloud Composer. For systems that offer connectors, our team wrote scripts in order to connect the tools with BigQuery. 

Finally, as our work extended so did their tool integration requests. Once we were done integrating one tool, they would add a new tool added to their portfolio that required integration. During these integrations we also were tasked with ensuring quality by performing regular data maintenance and quality tests.

The Result

A Data Platform on Google Cloud that centralizes and connects all data sources, brings complete control of the data to the client, and that uses recent high-quality data. 

Engineering a data warehouse architecture that would serve as the single source of truth for data analysis, insights, and marketing purposes was key. The solution was  to setup a Google Data Cloud with BigQuery. This would allow a unified 360 customer view and for departments to access data like API’s. 

To do this, we connected the current tools they were using into BigQuery so data could be pulled for analysis and then developed scripts to interconnect to other tools being used. Connecting these data sources to one central location would allow a complete data overview. 

To prevent the disruption of their current internal data workflows our team provided tool activation once on BigQuery. For example, we set up their Tableau server and tools such as Vertex AI where data analysts could write their own code to run data analysis as needed.

The Process

With the goal of maximizing results in the shortest amount of time, Crystalloids broke down the process into two stages: data collection & data implementation for consumption. For the collection phase we’ve set up an architecture and processes to collect data from  various tools and sources, like:

  • CloudSQL instances
  • Third party applications using API’s (e.g. Salesforce)
  • Flat files (CSV, JSON etc)
  • Data that is being published on Pub/Sub topics

Once extracted this data was then ingested into BigQuery by using:

  • Cloud Composer (Airflow)
  • Dataflow
  • Python Scripts

Once the data was implemented onto BigQuery, data was consumable by the following staff and departments: 

  • Data Scientists using Vertex AI
  • BI specialists using Tableau
  • Business Analysts

Technology Used

  • BigQuery
  • Cloud Storage
  • Cloud Functions
  • API Gateway
  • Compute Engine
  • Kubernetes
  • Cloud Composer
  • Dataflow
  • Cloud Pub/Sub
  • Cloud Datastore
  • Vertex AI

Share this story