|
1
|
|
|
2
|
- How much gas do you have left?
- How fast are you going?
- Do you need an oil change?
- What gear are you driving in?
- Is your turn signal on?
|
|
3
|
- Collection (finding all of the data sources)
- Integration (place data into a warehouse)
- Analysis (create analytical reports)
- Interpretation (show trends and periods)
- Presentation (easy, quick, graphical UI)
|
|
4
|
- Helps achieve the organizational goals
- Setup a process to monitor KPIs
- Marketing KPIs are most common
- KPIs need to be presented graphically
- KPIs can determine the data needed for the data warehouse
- KPIs are set by executive level management
- Quantitative, Practical, Directional, Actionable
|
|
5
|
- Customer related numbers:
- New customers acquired
- Status of existing customers
- Customer attrition
- Turnover generated by segments of the customers - these could be
demographic filters.
- Outstanding balances held by segments of customers and terms of payment
- these could be demographic filters.
- Collection of bad debts within customer relationships.
- Demographic analysis of individuals (potential customers) applying to
become customers, and the levels of approval, rejections and pending
numbers.
- Delinquency analysis of customers behind on payments.
- Profitability of customers by demographic segments and segmentation of
customers by profitability.
|
|
6
|
- Databases configured for OLAP use a multidimensional data model
- encompasses relational
reporting and data mining
- allowing for complex analytical and ad-hoc queries with a rapid
execution time
- Consolidate the organizations data into one data warehouse for
consistent reporting
|
|
7
|
- OLAP Analysis – Entails designing data structure for OLAP,
defining cubes and dimensions, design the MDX language for queries
- Data Integration – Primarily the ETL (Extract, Transform, Load)
components to move from OLTP to OLAP
- Reporting – End user reporting tools and integration with portals
and BI Solutions
|
|
8
|
|
|
9
|
|
|
10
|
|
|
11
|
|
|
12
|
- Setup a data warehouse in current database system, (Microsoft SQL
Server, MySql, Oracle, etc.)
- Setup a portal using Microsoft IIS or Apache
- Utilize the Pentaho BI Suite to handle the ETL process, Dimensional
Queries, Report Writing
- Find a partner that can help setup all of this
- Planning and management is very important
|
|
13
|
- Professional Open Source System
- Lower cost of entry and on-going licensing
- Has equal to or better features than Commercial BI Systems
- Local support of the system
- Use only the components needed
- Major companies using it with great success
- Transition to commercial system, only if needed
|
|
14
|
- Data Integration – Exchange of data from OLTP Database to OLAP
Database
- Reporting – Utilization of the Reporting Tools and Components
- OLAP Analysis – Design of a data structure to support utilizing
Cubic and Dimensional Analysis optimized for speed
- Dashboards – Executive level reporting and graphical components
presented in a logic grouping
- Data Mining – Conglomeration of historical data to derive trends
and profile the data
|
|
15
|
|
|
16
|
|
|
17
|
|
|
18
|
|
|
19
|
- OLAP Designer
- ETL Tool designer and implementation
- Report Writer programmer
- BI Server Admin
- DBA (To explain OLTP structures)
- Dimensional Query Language skills (MDX)
- Data Analyst to define needs
|
|
20
|
- Poor Data Quality
- Tools were too difficult for users to use
- Lack of appropriate skill set in organization
- Lack of management buy-in
- BI System was too complicated to deploy
- BI Tools did not meet the end users needs
- BI Technology was too expensive – ran out of budget
- Lack of professional services or support
|
|
21
|
- Implement the ETL Tool
- Select and get training for internal staff on Reporting Tools
- Design OLAP data model
- Design Reporting/Dashboard examples and turnover to internal staff
- Utilize Solutions4Ebiz at local BI rates as needed
|
|
22
|
- Propagating accurate, clean and timely data into the data warehouse is
of utmost importance
- Flashy graphical dashboards without data integrity is a huge problem
- Bad data means bad decisions
- Focus and test the ETL portions of the data warehouse before moving on
to reporting
- ETL can take a significant amount of time and tuning.
- Monitor, monitor and monitor the loading process
|
|
23
|
- Customer account behavior from a deposit and withdrawal perspective
- Customer response to marketing campaigns
- Aggregate on-hand deposits for certain time periods as compared to other
time periods
- Trend analysis of deposits, withdrawals broken down into methods such as
Checking, ATM and other.
- Commercial Loan ratios to overall deposits
- Commercial loan interest payments, principal paydown and lending limit
ratios.
- Trend analysis of individual customers whether commercial or residential
- Interest payout spread.
|
|
24
|
- Registered and accepted as a Pentaho Partner
- Certified Professional Status achieved
- Experience doing BI/Pentaho Projects
- Experience in Dimensional Modeling
- Setup our own Pentaho Server
- Can provide a prototype environment to prove concept
- Limit our involvement by setting up and training
- Experienced in the ETL process for data warehousing
- Can host the BI server, if needed
- Many years of experience with proprietary Databases
|
|
25
|
- Engage in the discovery process
- Important to focus efforts on a narrow project initially
- Strategize on the output of the report/dashboard
- Reverse engineer the data components
- Scope the potential data model
- Determine go/no go
- Go, probably entails internal training
|
|
26
|
- $8,500 per month spread out over 6 months
- 80 Hours of BI work per month
- Allows for training and product “absorption”
- Provides a predicted budget
- Postpone subscription until the end
- Objective is to involve staff in the BI project from beginning to end
|
Notes
