Determine what data from which sources is needed
This article is part of the series “Set your project up for success.”
Start with the assumption that this aspect is going to be difficult – even in successful projects.
Companies have a lot of unstructured data. Acknowledge that it will be a big task to get the "right" data in the "right" format.
Prepare clear documentation of what data you need, and consider all sides of it.
- In what format should data be?
- Where does it come from?
- How you will get it into AIMMS?
- In what order does it need to go in?
- How you will store it?
- In what type of database should it be stored?
- What will you do with the data after AIMMS has optimized it?
Consider how often you will need fresh data – and whether you need a full set of new data, or just a delta containing the changes to previous data.
The more details you know up front, the easier it will be to track down the data and explain to the responsible parties what they need to provide you.
Acknowledge that it will be a big task to get the "right" data in the "right" format.
"We're almost done, we're just waiting for the data." That's a line I hear too often, for too long, in many projects. And when the data ready, it turns out to be much bigger than the testing data that has been used during development and testing. Testing late with real data usually results in unexpected results, and major changes have to be made at the last moment before launch.
The sooner you can get "real" data to feed your model, the better. Testing with a small dataset is one thing, but at some point you need to test with the real thing – and plan extra time to accommodate once you have it, in case everything blows up.
The sooner you can get "real" data to feed your model, the better.
Data management and data transfer services are typically provided by the customer's IT department or by the partner. In some cases, a contractor may be needed. (What AIMMS supplies is described on pages 5 and 6 of the SLS.)
We now offer a managed service for integrating different ERP systems and other data sources with AIMMS applications - learn more about this new offer in our recent webinar.
Part 1, Devote time to key areas in your plan for success
Part 2, Make "Time to Value" a priority