BUSINESS INTELLIGENCE | DATA WAREHOUSING
Dashboards and data warehouses can be tricky things.
-
You can have so much functionality in a dashboard that a user might be seeing different data than they expect.
-
In particular, you need to make sure that every user action that affects the data being shown is obviously displayed to the user.
-
-
You can have so much information in a data warehouse that you must ensure you are obtaining exactly what you need.
-
Despite the organized and symmetrical data models in data warehouse courses, the actual data warehouse in a major corporation can be somewhat messy. The data from multiple sources has been imported into the data warehouse, but the I.T. staff might not have had time to do more than to just link it together. They have not had the time to restructure it.
-
Even if they have restructured it, it might be non-intuitive and you need to work closely with the data warehouse staff to ensure you are pulling the correct data.
-
Business Intelligence Definitions
The definition of Business Intelligence varies from source to source. These are the definitions I prefer. Business Intelligence is finding out what happened. Business Analytics is understanding why it happened. Prescriptive Analytics is predicting what will happen.
Using data pulled from the client's database systems, dashboards can easily show what happened and a user can use them to start to work out why it happened and what might happen.
Prescriptive Analytics is more an art than a science. The people who do this at one recent employer all have PhD's and look like wizards.
Business Intelligence
Most of the business intelligence work I have done has been to create QlikView dashboards. I have created dashboards for many different clients.
The typical engagement consisted of my finding out what the client wanted to know, gaining technical access to their database(s) and then creating one or more dashboards to provide the data they wanted. I would work with all levels of staff, ensuring each person's interests were satisfied.
The engagements were in a wide variety of industries. Financial, education, nursing homes and battery manufacturing are some examples. Two specific clients were DirecTV and Fairplex.
DirecTV
DirecTV has a large staff devoted to Business Intelligence and Business Analytics. They are continually working to provide more and better service. They are also continually working to identify their best customers and on how to keep them.
My job was to start them moving to QlikView and to do some of the most difficult dashboards and Teradata queries myself.
A big question for each dashboard created was "What have you wanted in the past that you had decided that you could not do?". As was typical of showing QlikView to new people, they were often amazed with what they could now do and how easy it was.
Fairplex
Fairplex is the non-profit organization that manages activities on a square mile of land in the eastern side of Los Angeles County. One of their activities is the annual Los Angeles County Fair, the county fair with the highest attendence in the country.
The fair lasts for three weeks. Fairplex has worked out a variety of ways the public can use that land and the equipment it uses during the rest of the year.
Fairplex has 17 different subsidiaries, the fair being one of them. Fairplex wanted to have the a dashboard that combined the P&L's of their four largest subsidiaries. The data for each company was stored separately and between them they used 3 different accounting packages. I imported all of this data into QlikView and integrated it.
Fairplex rents out their facilities for events. Drag racing, horse racing, exhibitions and swap meets are examples. For each event, Fairplex will provide space, food, parking and equipment. Each of these is provided by a different Fairplex company. Fairplex wanted to be able to see a combined P&L for each event.
To do this, I first had to convince the vendor of their equipment rental ERP to allow a cost center to be assigned to a P.O. line item, not just a P.O. I then had to change the Fairplex business processes to use this more detailed cost center and then had to read in this change to QlikView.
Korn Ferry
Korn Ferry has staff devoted to Business Intelligence, Business Analytics and Prescriptive Analytics. My job was to help build their new data warehouse.
The types of Prescriptive Analytics questions their staff answer are "based on the respondent's answers on an assessment, how good of a CEO will they be at a large corporation?". The staff who do this are very smart and this allows Korn Ferry to make very good placements of such positions.
My work at Korn Ferry was to assist in joining the different data sets brought into the data warehouse so they could be queried together. I later transitioned to QA work to ensure the new data warehouse matched the same data as the old data warehouse.