Getting Started with Data: The first book you should read to successfully get along with data. by Menegatti Gabriel & Team Simbiose Ventures

Author:Menegatti, Gabriel & Team, Simbiose Ventures [Menegatti, Gabriel]
Language: eng
Format: epub
Publisher: Simbiose Ventures
Published: 2020-11-20T16:00:00+00:00

5.3.4 Prescriptive analysis

The prescriptive analysis also deals with machine learning algorithms and is a logical step after a predictive analysis. Prescriptive analysis works to answer, “What should we do about this?” Instead of focusing on possible outcomes, like the predictions, it focuses on what other data might be related to whatever was found that might happen. It would then suggest actions that could potentially capitalize on opportunities or mitigate threats.

Imagine this: suppose you’ve run a predictive analysis to forecast your earnings for the next quarter. Using past data, a machine learning algorithm can suggest if your earnings will grow or not, and at what pace. But a prescriptive model will try to find what other data from your dataset might be correlated to your earnings and, if your objective is to grow 10% in the next quarter, it will indicate what actions you can take to make this happen.

Supposing you’ve fed this algorithm with data from your HR, Marketing, Sales, Software Development and Accounting, the algorithm could then find that the people you hired for the Data Analytics team strongly correlate to earnings from that period of time. The investment made on one specific marketing platform while increasing the number of posts weekly on social media could be correlated to earnings as well. These patterns can suggest that to grow 10% in the next quarter, you can hire two more software engineers and invest 15% more on a specific social media platform.

Of course, the algorithms will not present the outcomes like this, using this language. To make the calculations, every part of your data is transformed into a format that the algorithm can understand and correlate to everything else. The results need to be interpreted by someone used to these data types and transformations. And again, everything that emerges as a prediction needs to be carefully analyzed in the context of that business because the machine wouldn’t know who to hire to reach those outcomes. If none of the candidates have expertise in the technology you use, then hiring someone isn’t really the best option.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.