Transforming IT
Showing results for 
Search instead for 
Did you mean: 

The fast changing world of data & analytics


I was going to do one or more posts on how machine is, and can be, used in various industries.  However, before I do that, I thought I was take a quick, one-post, detour into how data and analytics is changing because that will, I think, help characterise the types of use of machine learning better.  

I think that the best way to talk about data and analytics has change is to talk about what we used to do and then what we are starting to do and will do. 

1. Real-time analytics

Used to happen

We used to use analytics look in the rear-view mirror – we used to analyse what had happened - “these were the sales”, “these were the support issues”.


We are now starting to do analysis in “real time” and then use this analysis to take action:

  • A self-driving car.
  • A drilling platform analytics solution that predicts when a drill bit will have problems and immediately stops drilling.
  • A healthcare system that looks at the “digital footprints” of patients and alerts that a patient may have kidney problems.

We call these “inference → action” systems, “Systems of Action”. 

2. Analysis of structured, human-interaction and IoT data

Used to happen 

We used to analyse just structured transactional data – sales transactions, customer returns, hospital visits.


We captured and processed “human interaction data” like texts, tweets and Facebook entries.

Now and in the future 

We digitize our analog world. This is what IoT (including video analytics) is all about. The amount of data we now need to process is huge. Walmat’s entire transactional warehouse is estimated to be about 6 petabytes. Facebook collects about 4 petabytes a day. Just one self-driving car will generate 4 terabytes of data in a day – just one car in one day.

Now, therefore, analytics is about doing analysis across all data types - structured data, human-interaction data and our digitization of analog reality.

So what? 

The amount of data is so huge and/or our Systems of Action require quick action and so we can’t send everything back “to the core” for processing. We need edge computing


3. Analytics inside and outside applications

Used to happen 

Analysis “without the app” – we looked at the data the app created once the app had stored its transactional data. In other words, our analysis was done from outside the app, looking into what users were doing with the app. 


Analysis is now also “within in the app” (and still “without the app”). We are moving from applications that are purely statements telling the computer what to do; statements of the form “if X then do Y”.

Our applications are starting to use machine learning (for a description of what machine learning and other types of AI are, please see this blog post). Rather than “if X, then do Y”, we ask the machine learning system to figure out what is going on, and then we use programming statements to take action.

For example:

  • a machine learning application used in healthcare will “learn” what the digital footprint of a patient with kidney disease looks like.
  • a machine learning system inside an application learns when an oil-drilling bit is going to have problems.
  • a machine learning system learns how customers likes to shop.
  • a machine learning system learns to predict when a locomotive is going to need maintenance.
  • a machine learning system learns what a certain type of cyber attack looks like. 

As we discussed in the post on AI and machine learning, machine learning has two components. There is the “run-time” inference portion and there is the training/learning portion. It’s the “real-time” inference portion that we embed inside applications.  So, rather than “if then else” statements, the app asks the machine learning system “what’s going on?”, and then it takes action depending upon how the machine learning system responds. 

4. The non-developer developer is getting the money

In a previous post, I talked about the rise of the non-developer developer - the developer inside the business who uses applications as a tool to get their job done. 

There are non-developer developers creating applications and, probably even more prevalent, there are non-developer developers create analytics - on sales data, on medical data, on IoT streams from oil pumps and locomotives. 

And so, we need to put in place the analytics platforms that these non-developer developer analysts need. These platforms need to be easy to use and almost certainly, they need to include all the latest Open Source because it’s thru doing better analytics than the competition that many businesses now gain their competitive advantage (more on this in the next section). For example, have you been asked for a system with TensorFlow on it yet? You will.

5. Data and analytics and its role in digital transformation

Businesses are increasingly using data analytics to gain competitive advantage. In what ways can analytics be used to do this? Some time ago, we in the “digital transformation” team at HPE created a model for how digital technology could be used by the business. Thru:

1. Better customer experiences

2. Better products and services or new products and services

3. Better back-end business processes

HPE’s high-performance computing group, who work with AI and especially deep learning, a lot, have amended this model and added two new areas to think about. Their model says that the business can use AI in the following five different business areas: 

1. To create a better business experience. For example, machine learning can be used to understand the needs of customers and therefore, to better personalise their customer experience. Or, as a heavy Spotify user, I look forward to Fridays when Spotify’s machine learning system will give me my very own, personalised, playlist of all the latest releases based upon my previous listening patterns

2. Better products and services or new products and services. Self-driving vehicles are stuffed full of AI

3. Back-end business processes (internal and internal to partners). For example, machine learning can be used to make supply chains better. Or, it can (and is) be used to identify counterfeit hotspots - supply chain areas that seem to be suffering from counterfeiting. 

4. Improving employee experience. I like this one. It seems to be that whenever there is a new technology we always focus on a/ how can we use it to make money and therefore, b/ customer situations. If you truly believe that your employees can give you a competitive advantage, then why not use AI to “increase their gearing” and allow them to do the stuff that is interesting, cutting out the boring and irritating stuff. For example, I’ve seen many a sales portal over the years. Each one looks nicer than the last, but when they are populated with a few years’ worth of presentations, sales guides, battle cards and the like, they all become very difficult to use - it’s almost impossible to find what you are looking for. Why not try using both AI bots to help with the search for material and machine learning to characterise how each sales rep like to get his/her information?

5. AI will be used to create completely new business models. Look for intermediaries to be completely cut out of the loop. For example, in the UK there is a machine-learning based system that writes your will for you, cutting out the lawyers. There is, of course, debate about whether or not this system is as good as a lawyer, but it’s certainly a lot less expensive, and if it’s a good machine learning system, it will be getting better all the time - machine learning systems do that !

So what?

I personally find models like the one above very useful for brainstorming and planning. You can go thru each of the five areas and ask yourself, “how could we use AI for this category?” 

Also, if you have someone in your organization who is starting to track what is happening in digitization in general and AI in particular (and I suggest that you really should), they could categorise what they find against the above model.

Related posts

Mike Shaw
Director Strategic Marketing
Hewlett Packard Enterprise

twitter.gif @mike_j_shaw
linkedin.gif Mike Shaw

Mike Shaw
Director Strategic Marketing

linkedin.gifMike Shaw

0 Kudos
About the Author


Mike has been with HPE for 30 years. Half of that time was in research and development, mainly as an architect. The other 15 years has been spent in product management, product marketing, and now, strategic marketing. .


Thanks for sharing. Any thoughts on how D&I could/is transforming Info Sec?

Read for dates
HPE at 2019 Technology Events
Learn about the technology events where Hewlett Packard Enterprise will have a presence in 2019.
Read more
Read for dates
HPE Webinars - 2019
Find out about this year's live broadcasts and on-demand webinars.
Read more
View all