We recently discussed how a logistic regression can be used to model a binary response and predict the probability of a binary outcome. However, there are instances when we need to take this a step further and classify the outcome based on a cut-off probability threshold.
read moreRetailers have been finding more and more sophisticated ways to use customer data to their advantage. By collecting and analysing consumer data, together with other socio-economic data, supermarkets are able to make evidence-based decisions when devising their marketing and operational strategies.
read moreSurveys often comprise tick-box questions, resulting in what are referred to statistically as categorical data. Analysing categorical data requires the use of a specialist set of statistical tools as they are not normally amenable to the standard tools available for continuous data.
read moreAccording to IBM's calculations, "every day, we create 2.5 quintillion (2.5 billion billions, or 25 with 17 zeros) bytes of data! Not only are new forms of computer processing being developed to handle this so called 'big data', but the way we are communicating data is also evolving.
read moreWith the majority of large commercial organisations embracing Big Data as a means to refine their products and better understand their customers, the charitable sector has been slow to adopt the same techniques to streamline their operations and increase their donation rates.
read moreInteracting with social media is a key element of any strategy for businesses wishing both to better understand and interact with their customers. But how can you measure the success of a marketing campaign on Twitter?
read more