THE ESSENTIAL BUSINESS INTELLIGENCE
R&D² – Research, Discovery, Development
Effective learning – it all begins with research (understanding the context, digging out the evidence), discovery (uncovering insight through forensic analysis), and development (creating, implementing, evaluating solutions).
This three-part series lifts the lid on using data with machine learning technologies to drive business intelligence and insight. Part one looks at sourcing data – not as easy as it sounds. Part two focuses on preparing data, actually the biggest and most complex part of the whole task – and likely where the biases and errors creep in. Then in part three, we lift the lid on the engine and take a look at machine learning in action.
BUSINESS INTELLIGENCE – Part one
Sourcing your data
‘Business intelligence’ (BI) is a pretty confusing topic. That is not helped by its being the subject of several slightly different definitions. What they all share in common, though, are two features: first, BI is about data, and the application of tools to make data reveal enlightening and informing truths so that decisions may be better informed. Second, they tend not to make any reference to where the data comes from and how it has been acquired or captured. So, according to these prescriptions, data is somehow conjured, in a form that can be applied to some task, and that with the right tools, data will be manipulated to provide insight sufficient to inform management decision-making.
That’s the brochure description. Now here’s the real one: it is – or should be – about a whole lot more. BI starts with a problem, challenge, issue – call it what you want – some feature of a business that needs some kind of attention. We can call it the hypothesis-space. The initial priority is to objectively evaluate the validity of that ‘hypothesis’. A hypothesis is ideally quite specific and focused. So, there might need to be several hypotheses. In the main, we are challenging assumption, perception and subjectivity.
BUSINESS INTELLIGENCE – Part two
Preparing Data for modelling
Business Intelligence (BI) is an essential tool for making decisions and judgements that are better informed, and which can lead to improved business performance. It is a rigorous scientifically-grounded discipline which uses good research practice combined with often highly sophisticated computational tools to generate meaningful, accurate and actionable insight from data.
As we saw in part one of “Understanding Business Intelligence” [LINK}, sourcing data is in itself a potentially complex undertaking scattered with potential pitfalls. We also made the point that ‘rubbish in, rubbish out’. In this second part, we dive into the topic of data preparation. To many this may seem frightfully boring and best left behind closed doors but actually it’s the meat and potatoes of the whole enterprise.
UNDERSTANDING BUSINESS INTELLIGENCE – Part three
Generating Business Intelligence
Using Business Intelligence (BI) effectively brings many advantages: empowering executives and managers to make better informed decisions, for instance, or helping to cut costs, or highlighting processes or practices that could be improved. In fact, if you subscribe to research and thought pieces by organisations such as Gartner, BI is transforming not just the way organisations think and act, but also how they act. So it is little surprise that BI is attracting a great deal of attention and interest, with an increasing number of BI off-the-shelf tools and platforms for organisations to dig into.
Establishing BI at the enterprise level is not an easy task. One large College that I worked with has taken the best part of a painful year to implement their chosen BI solution: now it’s up and running, it’s impressive and delivering the kinds of analytics results they need. These tools are mostly designed as end-to-end solutions, which may or may not be powered by AI technologies.
Get in touch
Professional. Expert. Limitless curiosity and ambition to make a difference.