Custom Process Management DB
Started a project to build a custom MS Access database for one of the machineLOGIC clients. They need a simple way to track tasks related to processing employees when they start or when they leave. Access was deemed the best way to build them something quickly and inexpensively.
The database design is just a handful of "core" tables in 3NF and half a dozen lists tables used in the Access forms for combo boxes and such. There are main application forms along with lookup date management forms where they can mark items active/inactive and control the display order. While Access isn't always the best choice, for an app like this with just a couple of users it is a great fit.
As I write this (early May) we're just about done wrapping up the reports part of the project which is always the most iterative or interactive (imho).
Data Conversion / ETL Project
Spent the last 2.5 months of evenings working on project for several (inter-related) machineLOGIC clients. They're all moving to an online web application from Zywave and they all have legacy data in a variety of formats. That legacy data came from their two existing applications: CDS and Goldmine.
To keep it interesting, the implementations of CDS and Goldmine were rather customized and very flexible. That definitely caused some extra challenges when doing the conversion.
Zywave helps make it easy for conversions though. They provide a set of spreadsheet templates and if you somehow get your data into those templates they can get it loaded. Thus, the challenge was getting the data out of the current systems and then massaged into a useful dataset that could be dumped into spreadsheets.
Enter MS Access 2007. Access makes linking to external data a breeze and the (reasonably) standard SQL implementation makes it a powerful tool in a low-buck ETL project. And that's exactly how it worked for me on this one. I'd pull the data out of the existing systems into CSV and then link those CSV files into Access to look like tables. Then I'd run my queries and generate data suitable for exporting to the templates.
This project went several months, mostly due to the amount of data cleansing that had to be done before we could send it to Zywave.
