KFC, a subsidiary of Yum! Brands, Inc., is a global chicken restaurant brand with a rich, decades-long history of success and innovation. Today, the company has more than 25000 restaurants in over 145 countries around the world.
Elaboration of the Corporate Data Warehouse
One of the KFC`s representatives faced the following problem: the current data warehouse system (DWH) was based on foreign software (Qlik Sense), which involved the risk in terms of further development and support of the system in its local area. Due to the processes of localization of the corporate accounting and BI-reporting systems the company had a need for the development of a local corporate data warehouse (CDW).
The project is aimed to create a new analytical landscape and consolidate data from multiple sources in a single place (CWD), unify data and use the corporate data warehouse as a single source for generation of the analytical reports and other different consumer systems.
Within the CWD elaboration we connected to a great variety of the data sources and implemented two data layers: raw data layer dt1 and pre-calculated data dt2.
The connected data sources are the following systems:
- RKeeper;
- Axapta;
- Other sources:
- BOSS
- IRB Reports
- Digital
- OFD
- XLS-files
- Omnitracker
GreenPlum was selected as the DBMS.
Apach NiFi and Apach Airflow were used as ETL tool.
Grafana was used in order to monitor the data, Node Exporter was used to collect the metrics within the server and Prometeus was used to collect the metrics from each server’s Node Exporter.
The result of the CWD elaboration project was the implementation of a fault-tolerant GreenPlum cluster with two data layers and configured automatic replication, containing data from all the sources listed above.
In addition to that, we implemented automatic data update in accordance with the individual regulations for each data source.
Additional measures connected with the optimization of the data warehouse:
- Table balancing;
- Partitioning (monthly and line-by-line);
- Distribution of tables by segments;
- Changing compression type;
- Vacuuming;
- Script analysis and optimization.
The generated logical model of the detailed level contains the data necessary for further development of data showcases (layer dt3), which will provide the possibility of both creating reports and using almost any BI-tool of the Client’s choice.
The developed CWD provides the possibility of expanding the attribute composition without a significant increase in the cost of the solution, i.e. it does not lead to uncontrolled growth of costs when adding new business entities.
Despite the achieved significant results we are still working at the further improvement of the analytical tool.
Technical support of the Analytical Reporting System
Within the signed Technical Support Agreement the Client’s analytical reporting system based on Qlik Sense is still supported.
Our specialists process the Clients` requests, which are published on the GSD portal, in 5/2 mode from 6:00 am to 18:00 pm. Processing of the requests is carried out within the SLA which contains the classification of requests by their criticality, the speed of response of our specialists and terms of an incident elimination in accordance with its criticality.
Development of the Analytical Reporting System
Another type of our cooperation with the Client is the development of the existing Analytical Reporting System based on Qlik Sense.
The following modifications should be mentioned:
- Elaboration of new reports;
- Modification of existing reports;
- Correction of errors in current reports beyond the scope of technical support;
- Development or modification of NPrinting reports for mailing.