Data presentation and visualisation
No matter what level or type of data analysis, Machine Learning (ML) or Artificial Intelligence (AI) we use, the results are no use if they are trapped in a computer’s memory. There are two broad method of using this analysis: we can allow autonomous actions, or we can supervise these actions by accepting or proposing decisions manually.
Autonomous decisions are fairly widespread. For instance, think of simple tasks such as stock re-ordering. This could be achieved by a supervised technique whereby pre-defined minimum stock levels decide whether a product is classed as ‘sufficient stock’ or ‘requires ordering’. Additionally, unsupervised techniques may learn and refine the minimum stock level to adapt to changes in delivery times, for example. But what if product sales are too low, do we want to continue to stock this item? In some cases we may wish to drop these products. In other cases these products may impact on others, eg, ‘loss leaders’. By using appropriate visualisation techniques we can use ML & AI to identify these products, but it is equally likely that human intervention would be required before cancelling this item completely.
To aid decisions such as this we need to present our analysis clearly and we may need a range of techniques so that we don’t just look at sales numbers, but also the inter-relationship between products.
If you sign up for the ‘AI Technologies for Business and Management’ program, the fourth short course ‘Applications of AI in Business’ will cover data presentation in more detail.
© Coventry University. CC BY-NC 4.0