Rodney Matejek

I always believe that any development work has to include good analysis work as well.  Recently I was working on a project that focused exclusively on data analysis, and here is how I approached my work.  

 

It was an assignment for the public sector industry, where there were data feeds coming from county municipalities throughout all fifty states.  The data feeds were coming in every fifteen minutes from each site every day, seven days a week. Here are the considerations we had to account for as we did our work.

 

1. Each locality had the choice of as many as fourty different software vendors to choose from when modernizing their processes.  We had to contact the software vendor to discover how their system was designed to work to get a head start on understanding how the individual counties processed their information.

 

2.  We had to contact the individual counties to discover if they were using the software as the vendor designed it to be used, or if they adopted a procedure that would change the way data was being saved.  We also knew that the counties may not have a strict data administrator on staff, and so may not understand standard data processing procedures.

 

3. We would remotely log onto the individual agency's servers to look at the data ourselves to find out how they were collecting and saving their own data.  This would require giving each county a laptop, that they would connect to their server, and would allow us to see it for ourselves.  If there were privileges that were needed to be granted to our laptop in order to get to correct data, then negotiations would begin to find out how comfortable the vendor would be to grant us access.  We would have to work with what we had.

 

4. Statistical information was derived from examining the data to see the following.

   a. How the data was formatted in the database.

   b. If there was any data dictionaries available to use in code translations.

   c. To see how often reliable data was being collected into the system.

   d. Data mappings to be used when collecting the individual county's data, and how it would fit into our     own database.