Let's start with the bad news: There are no magic solutions here. Data connectivity is an aspirational goal, like operational efficiency As defined in this article, "data connectivity" is the act of taking information from different computer systems and combining it to gain better insights. As you have probably seen already, that's harder than it sounds.
Why is data connectivity so difficult? It makes sense that computer systems should be able to talk to each other. And in fact, they can. There are endless ways for computer systems to share information, starting with the prosaic .csv file (.csv = comma-separated values) and going all the way through real-time direct access. The problem is the data.
You are probably familiar with one way in which systems talk: When one system imports or exports a file of information from or to a second system --an "interface," as we information technology (IT) people like to call them. Financial systems have lots of interfaces. They import files for posting to the general ledger and export formats like bank files or 1RS layouts. (All the examples in this article will be for financial systems, but the same ideas would apply regardless of the data involved.)
Data connectivity is slightly different from an interface; it's the idea of pooling data from two or more systems so you can ask questions about the combined information. Think about it as two different systems that are both exporting information to a third location, which is a reporting database, where people can run queries against it.
The challenge of data connectivity is finding the commonality between systems. How do you make a financial system "speak" to a property database? Or police arrests? They usually have different transaction formats, reference codes, and even inconsistent code values to represent the same thing. (There's an example later with organization codes.) The rest of this article explains how to approach this problem, although of course actual results will vary because every situation is different.
Before we begin, some terminology. Computer systems generally have two types of data: transactional data and reference codes. Transactional data are the detailed history of events, which are classified using the reference codes. Financial systems are full of transactional data: general ledger postings, invoices (paid or billed), budget requests, etc. Reference codes are the pre-defined values that group transactional data such as object codes, organization codes, vendor codes, funds, grants, etc.
STEP ONE--KNOW YOUR DATA
Connecting data starts with an understanding of what's there and how it's stored. Actually, let's back up--it should start with having a business question that's worth the effort. "Let's just mix all the data together and see what we find" is...