Data Warehouse & Data Migration

  • Design database for large scale system (Multi talent and Complicated data)
  • Migration and Synchronized solution
  • No-SQL Solution (HBASE, Hive, Cassandra, Graph DB)
  • Flexible Data Warehouse Design with flexible ETL
  • Build MDX OLAP Cube
  • Big data design with Hadoop 2 Ecosystem
  • Datamining and Machine Learning implementation

One of the main parts for enhance the quality of report toward building of BI system is to build up a centralized data store for your organization. Below are the main steps of analysis and development approach:

  1. The first step, we should collect all data sources including database, file, .csv, Api… that involve the current data process.
  2. Investigate the all data sources and determine what classification/categories data are sharing for multiple application? What data need to keep historically? Compare and make matrix for them.
  3. Review all NT’s reports and compare these report fields with the current data to determine as much as its information and knowledge to understand its current business data.
  4. After having these iterations analyze data and reports, we should do more technically with the design standardized data model that will store all data from data source. This step is really challenge when standardized model must be adapted and remove the duplicate and redundancy data.
  5. When having data warehouse model we should develop the way data feed into it, in the technical concept, we have to deploy ETL transformation to mapping from original data source to new data store model. Especially we need to include the calculation, business rule when transform data.
  6. Next step looks quite complicated with tons of expert knowledge on data warehouse & DataMart process.
    • Design Data warehouse and DataMart
    • Design DWH model
    • Master data/ hierarchies
    • Define Dim & Fact data
    • Define Schema
    • Historical data
    • Separate Subject Areas
    • Identify the grain
    • Define security role
  7. After collect, transform, aggregate, summarized data by speared Data mart we should do next steps to deliver the reports, dashboard, and discovery knowledge (mining, forecast data) as:
    • Design Multidimensional Model
    • Implement OLAP cube
    • Choose Mining algorithms
    • Implement Predicting Model
  8. The final step we only need to display BI data on the user-friendly visualization as Power BI, Tableau, Qlikview.

 

Below is some Case study we have built for our client


  • Abiding Technology Investment
    Head Office: #35 Block C1, Dai Kim-Dinh Cong New Urban, Hoang Mai, Ha Noi.
    Branch: Room #6, 2nd floor, Sevice Office, Truong Thanh Building, Trang An Complex, No1 Phung Chi Kien, Cau Giay, Ha Noi.
    Telephone: +84(0) 90-247-1166/ +84 (0) 868 463339
    Email: contact@abidingglobal.com
    Skype: Abiding Tech
  • Products: 2TCommerce