'Bad data in - Bad data out'

- August 26, 2020

For those veterans that are still wondering what data analysis is and what role they can play! Here is a quick story to get you read in.

Veterans out there probably look at the chart below and wonder...what are these people talking about?! Every time I talk to someone new to the space they say - "Data Analyst - that is like Artificial Intelligence right." The answer is - 'not really - the machine learning and AI algorithms are only a small percentage of data analysis.' You can't enable the top of the pyramid if the foundation and the lower rungs are not constructed properly. I will explain....

A very complicated picture.

In the Army, you need to organize how you bring in data to the decision making process. If the section commander needs to make a decision, he needs to collect it - shots fired, 'what was that', ask questions and demands 'send situation report' - move/store it - write it on their map, send by radio - transform it - prepare a contact report and/or medical evacuation request - aggregate it - make a decision, call for help - learn from it - After Action Review - optimize it - run the scenario in a simulation 1,000 times and see if there were better options - this is also done through repetitive, challenging and interesting training.

So what? If you are collecting the wrong/inaccurate information, not storing it properly, preparing incomplete or faulty reports, that lead to bad decisions that support zero learning - good luck optimizing. 'Bad data in - Bad data out.' You need to properly collect, store, transform and aggregate to make good decisions, learn and optimize.

Computer Errors: What can you do?
After pushing the optimize button on Artificial Intelligence using 'bad data in - bad data out'.

There is a risk when we all move toward the shiny AI and machine learning solutions and think that it will solve our problems. Small Wars Journal points out that, "The real concern, though, is that military leaders may not comprehend significant risks associated with blindly using such tools." Getting a computer programmer to feed a bunch of data into a machine learning tool - while asking 'what's a battle?' is not going to optimize the problem. That is because, they don't understand the problem. This is where you - the veteran - is critical in government, business or military to solve problems.

Further, as Deloitte points out, there is a concern of data overload. Sending the section commander 10-figure grids by voice, the color of the sky and everyone's favorite sports team during a fire fight is not going to help. As Deloitte points out, Commanders want to know “How does any of this information help me understand . . . what [operational] decisions are needed? Most of this is just information without analysis.” They will need your help. They need an analytics translator.

You are able to use the four step process below to understand the data. You define "I want to know where is the enemy hits the section?" - transforming the data by getting grids, contact reports, interviews with the section to understand their perspective - you analyse what you have and communicate your findings. As of today - a computer can't do this. They need real world problem solvers to investigate and ask the right questions. That is where you come in. As Deloitte points out in the section "Anticipating enemy fire: Mission analytics for threat assessment," if you do data analytics properly and define the right questions, you can start to know where the hostile actions will occur before they happen. That is what you would provide the company/government/military when trained. You are working out what information is important to the organization to make the right decisions - preferably before the firefight.

WYWM Basic Excel and Analysis Skills - Common Errors - Rallypoint

What is your role in Data Analytics? It is asking the right questions, seeing what data is out there and then asking more questions....WYWM will give you the tools to help transform, analyse and communicate - but the tools are useless if you don't define the problem. If the section members are feeding the Section Commander 'bad data in,' then he will be forced to give 'bad data out.'

Working in Kandahar Afghanistan - helping everyone do data analysis...


If you want to break into the tech industry then sign up to our platform and begin your training today.

Leave a Reply

Your email address will not be published. Required fields are marked *

Join our community

We have a Discord server where you’ll be able to chat with your instructors and cohort. Stay active in your learning!
Join discord