The Vaccine Distribution Cold Chain, Hungry Feeders and Sawing Logs (the data kind, not the tree kind ;-))

In the coming weeks and months governments all over the world will attempt to inoculate ~50% - ~70% of their citizens who are dispersed across every village, county, state, province and country on earth. 

Flowing through those 1 millimeter thick hypodermic needles will be 0.3 ml doses of supercooled vaccine that had just been produced by a growing list of biotech firms. 

If we work backwards from the arm that needs to get inoculated, through the cold vaccine supply chain, to the originating biotech tech firm that manufactured it, you can imagine countless variables that need to be considered. 

In normal times, this type of complexity is tamed by employing equally complex data models that are built well ahead of time and are usually based on well tested prior models. I think it is fair to say that this effort will be different. 

Tens of thousands of data architects, developers and analysts will collaborate to build thousands of local and regional data models. Those models will need to roll up to help plan the global distribution efforts. These efforts will span everything from how many doses did the UK purchase, to how many doses are need to be delivered to each village in Spain, to the timing of deliveries at 1 of the 3 sites in Springfield Illinois, USA. 

Many data platforms will be used in this effort. I live mainly in the world of IBM Planning Analytics (a.k.a. TM1) these days which is particularly good at modeling complexity and performing what-if analysis and scenario modeling, fast! 

So, what do these three things: Covid-19 Vaccine Cold Chain, TM1 Feeders and Logs all have in common? 

They are all required to insure that the data models - big and small - are designed, developed and tested for accuracy, performance and completeness. But why specifically did we jump from global vaccine distribution to TM1 feeders? Because, just like the last mile of the Covid-19 Vaccine Cold Chain is critical in getting people vaccinated, accurate and complete data in the data models could be considered the first mile of that chain.

These models will contain enormous amounts of metadata and they will need to be fast and flexible. So, if IBM Planning Analytics (TM1) is the platform of choice then lots of feeders will need to be built, tested and debugged with the requisite methods to keep these models performing. A slow down or miscalculation in the first mile could have enormously amplified effects at the last mile.

• Covid-19 Vaccine Distribution Cold Chain: This refers to the need for a “dose” of covid 19 vaccine to be kept “cold” continually from it creation in the factory to its delivery into the recipient. 
• TM1 Feeders: FEEDERS are used by the IBM Cognos TM1 calculation engine to assist with handling sparsity in cubes
• Logs: More specifically the TM1 message log records status messages to and from the TM1 server.

So, why am I bringing these ideas Cold Chain, IBM Planning Analytics (TM1) and Logs together? 

Simple, really. To champion the need for "attention to detail" as early in the process as possible. Whether it is global cube rules, MTQs, feeders etc...

It is in these early design decisions that planning models will have their greatest effects on the overall system and process outcomes. Taking every step to engineer errors out of the system early will protect against cascading failures like production shortfalls, distribution bottlenecks and vaccine spoilage. 

For years now I have worked with many of the the world's largest companies to support the accuracy and management of their TM1 models and if there is one thing I'm convinced of it is that if the problem is big and complex, the details matter, a lot!

If you would like more information about IBM Planning Analytics or automated model metadata management tools like QUBEdocs please reach out to me via LinkedIn Messenger or at our website below.

Scott Felten, GM, QUBEdocs