Data Conversion / ETL / BI Developer
If you need someone to fix an engine, you won't go wrong hiring someone who's actually built one. Joe can easily utilize your organization's Extract, Transfer and Load (ETL) application because he's written his own library of ETL tools.
Few are more experienced
It goes back to Joe's first year in the software business, when welcoming a new customer meant converting the organization's database.
Nowadays most databases are rendered in some flavor of SQL and therefore are easy to read and distill. In the 90s there were at least a half dozen database formats commonly in use (anyone remember Btrieve or Paradox?) but few methods to reformat them. So Joe and his colleagues wrote their own software to glean data from foreign formats. Using these tools, Joe has converted hundreds of new-client databases over the years, including one year when he converted the systems of about 100 veterinary clinics.
The process of cleaning up data such as dates, numbers and names requires a separate set of tools and skills. Here is where Joe stands out. In addition to name- and number-reformatting, Joe has written software -- and "tweaked" it over the years -- to parse names and addresses, standardize addresses, resolve foreign postal codes, and selectively capitalize (i.e. "USA" instead of "Usa", "Blvd" instead of "BLVD", "O'Connor" instead of "Oconnor").
The artfulness of the extraction and transfer process counts for
naught if, after everything is loaded, the reports don't balance or
if two mailers arrive at the same household. Joe has
conceived and developed a variety of software routines to ensure transactional integrity and manage duplicates. A few of
these are described below.
A sampling of ETL and Business-Intelligence (BI) projects
Developed an application to match a name-and-address database with census blocks, using the street- and address-range information from a database offered by the US Census Bureau.
Enabled users to import their own CSV files, matching fields with a drag-and-drop function. This is one of the functions included in the Filpac demo.
Enabled new users to access their own database in the menu-driven Filpac system by importing years worth of their own filings from the Federal Election Commission's system.
A synchronization routine for those users who wish to maintain virually the same database but on separate networks and PCs.
Duplicate-processing features enabling users to quickly scan, spot and merge duplicate records in databases built from many sources. This led to the creation of a donor-prospect database (three million transactions and counting) from hundreds of separate public filings.