The Usefulness of Graph Analytics in Data Science Projects
Date: January 21, 2019
|
Posted by:

Situation Description

Data these days is complex and while there are numerous ways to make sense out of it using Stats and Machine Learning, oftentimes this doesn’t suffice. Also, there are many times that we need to understand the relationships among the entities represented in the data, something that is not straight-forward in the majority of cases.

Problems Arising from This Situation

It is known to cyber-security experts that a mediocre system is worse than no system at all, since it just brings about a false sense of security to its users. Same goes with data science systems. Having a mediocre data science system to process a complex dataset is worse than not having anything at all, since the mediocre system is bound to either yield mediocre insights and/or fail to convince you that your data is of any value (even if the insights it yields are interesting). Since interesting insights don’t always equate to business value, special care must be taken when dealing with such data, data that can usually lend itself to a more in-depth analysis through a more specialized approach.

Implications of These Problems

As these problems remain not addressed, they are bound to have noticeable implication to the organization, resulting to additional costs and loss of potential revenue. For example, if restocking were to be done at a time ensuring that no item would be out of stock (or that no item would be overstocked), this would result to additional sales. In addition, if social media promotion was done at the days / times of maximal impact, it would bring about more engagement from potential customers, and therefore more revenue in the long term. However, by leaving these problems unresolved, they are bound to limit the profit, while potentially also damaging the dynamics of the organization’s ecosystem.

Enter Graph Analytics

This fairly niche methodology of data science, that’s based on a well established branch of Mathematics called Graph Theory, is at the epicenter of most advanced data analytics pipelines. Without being excessively complicated or too math-heavy, it manages to explore, explain, and enhance your data science processes, when used properly. After all, just like any niche know-how, it requires a certain level of expertise in order to apply effectively and efficiently.

Graph Analytics can help shed light on complex datasets where several customers/users are involved, who not only interact with your business, but may also interact with each other or with 3rd parties. These interactions are modeled as graphs (not to be confused with the plots used in data visualization), highly intuitive representations of relationships in a simple manner, without significant loss of information. However, graphs are in a dimensionless space, so creating them in a way that they reflect the original signal takes some effort. Fortunately, in every data science language, this whole process is greatly facilitated and to some extent automated. So, if you have someone in your ranks that understands this methodology and is able to handle the corresponding programming libraries, using all that to analyze your data in depth is the next logical step. As a bonus, you can see some interesting visuals of the corresponding graphs, so that you can get at least an intuitive understanding of what your data science professional is doing.

Next Steps

We at Data Science Partnership are able to offer various solutions regarding this methodology, customized to your organization’s needs. Be it by supplying the right Graph Analytics expert for you, or handling some projects ourselves, through a consultancy project, we can add value to your business and enable you to make the most of this very promising and still niche methodology. For more information, feel free to contact us today: cwright@dsp.ai, or call on +44 208 133 0822

Share with...


Zacharias Voulgaris

Zach is the Chief Technical Officer at Data Science Partnership. He studied Production Engineering and Management at the Technical University of Crete, shifted to Computer Science through a Masters in Information Systems & Technology (City University of London), and then to Data Science through a PhD on Machine Learning (University of London). He has worked at Georgia Tech as a Research Fellow, at an e-marketing startup in Cyprus as an SEO manager, and as a Data Scientist in both Elavon (GA) and G2 (WA). He also was a Program Manager at Microsoft, on a data analytics pipeline for Bing.

0 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *