The beauty of moving everything to the Cloud is that you have it in one place. For many, the start of this is the journey into data collection and storage, where we start putting all the data together.
Once you've done that you can begin analysing the data and building out all sorts of models. Doing it in the cloud means you have a lot of freedom to experiment. You can take the data you have accumulated through catalogues, data warehouse, and data lakes. You can make a quick prototype to see if something is going to float or sink.
The most successful people making use of their data take advantage of the freedom and flexibility that you have got that brings innovation and success.
There are a lot of data services available to help you get the most out of your data once you've moved to public cloud. The hot topics right now are AI and Machine Learning (ML).
Machine Learning is the concept that a computer program can learn and adapt to new data without human intervention. It's used anywhere, from automating tasks to offering intelligent insights, and industries in every sector are trying to benefit from it. You may already be using a device at home that utilises ML, like Amazon's Alexa or Netflix suggesting movies we might like to see next.
There's a wealth of data services that are available. At the basic end, simple analytics can help you look at your historical data to see what happened. On a more complex level, real-time data can help you carry out more predictive analysis to see what could happen and why something should be done.
“Being able to forecast into the future is a big part of what makes a business profitable.”
Let's break down the complexity of Machine Learning Services on three tiers:
At this level you don't have to do anything with your data. You can simply make use of off-the-shelf services. For example, extracting text from documents, or categorising images. You can start doing fundamental analysis straight away.
One level up, you need to add some ingredients, which is your data. The beauty of this is that there is a selection of different features you can then put in place. Fraud detection is one example. Personalisation or recommendation systems are another.
You can experiment with these by extracting your data, formatting it in a particular way, and giving it to these services to start getting actionable insights right away. You can also hook into a streaming analytics pipeline so you can start taking data from other services and have it flowing into Public Cloud, helping to make these predictions.
You’re now at the MasterChef end of the line. If your data has enough in it that you can drive the insights you want, and there are good predictors in there, you can go down to the 'raw ingredient' tier of services where you can do the deep data science. Examples of this are face recognition and self driving cars. By building the model yourself you can tailor it perfectly to your needs
The "ready-made" and "part-baked" services mean you can see whether there is something that is going to work for you rather than commit upfront to months of data science that might not produce the model that's going to be of use to your business. It also means that you can begin building the supporting infrastructure while you create the best model for your business.
The tiers show us about the amount of your data and effort required to what kind of impact the outcome gives. In essence:
Ready-made = Less effort, less of your data => less impact
Part-baked = More effort, more data => more impact
Raw ingredients = Most effort, most data => most impact
In the past, only the mega-brands with big budgets could take advantage of AI. But let me assure you - it's now for everyone and Machine Learning is making the art of predicting more affordable. Being able to forecast into the future is a big part of what makes a business profitable. It no longer takes a massive investment to run advanced algorithms and receive answers to a variety of different operational questions. With this knowledge gap being bridged, smaller companies have more power than ever before.
To get started with any kind of AI, look at where you stand today. What big problems need to be solved? What type of infrastructure do you have? What talent gaps exist on your team? While automation is the name of the game with ML, there's some management and hands-on work necessary to keep things running smoothly.
Always keep in mind, making too many significant changes at once will make it harder for ML to become fully embedded in your business. For most organisations, implementing ML will be a slow and steady process that will form a return on investment well into the future.
Think slow cooking, not fast food.