Is Augmented Analytics a solution to the data scientists’ gap?
Posted by Ramla Jarrar on Dec. 20, 2018
The exponential growth in data generated around the world has made Data Scientists one of the most valuable currencies in the job market, but probably one of the most expensive assets for organisations to acquire.
Conversely, according to Gartner “By 2025, a scarcity of data scientists will no longer hinder the adoption of data science and machine learning in organisations”
What is then the miracle solution that will allow organisations adopt data science and machine learning despite the widely known shortage in ‘Guru’ data scientists?
Could adopting Augmented Analytics powered by “Citizen data scientists” be the solution to restore the balance in a world more and more keen in adopting Data Science in each Functional Area of Business? Could Augmented Analytics be the way for a wider Data Science adoption across Mature and less Mature Organisations?
So what is Augmented Analytics?
Augmented Analytics are solutions that:
- Automate data exploration and provide relevant statistics without requiring deep domain knowledge,
- Create easy visualisations,
- Build advanced analytics models: Algorithms find patterns in data, auto-select features/Variables/Models that are best suited to datasets (e.g. Genetic Algorithms),
- Provide predictive models without a single line of code,
- Deliver prescriptive analytics to end-users.
It is basically automating machine learning and data science to do the ‘dirty’ job that would otherwise take 60% to 80% of typical analytics’ project timeline. In other terms, it enables the removal of bottlenecks that data scientists keep facing when performing analytics’ projects. This will save substantial time that could be reinvested in digging further into the results and devising better recommendations to the business.
The legitimate question that one may ask when introducing the concept of augmented analytics in the context of remedying to the shortage of data scientists is:
Can augmented analytics be an alternative to hiring data scientists?
The answer is yes if it is powered and led by a Citizen Data Scientist but let’s first define the concept of ‘Data Scientist’.
A Citizen Data Scientist (CDS) is someone who can create predictive and prescriptive models without necessarily being in the organisation’s analytics department. CDSs are not expected to be experts in statistics, machine learning or IT, only a broad knowledge of theses domains enables them to produce advanced analytics using augmented solutions.
CDSs are cousins of Business Analysts, only some weeks of training and introduction to the main concepts of the field can transform the latter to the former.
The advantage of CDS is that they are imbedded in the business and are therefore aware of the daily challenges faced by the business and the change in the emergency scale when it comes to prioritising business requests.
Hence, augmented analytics are likely to offer them the required flexibility to use machine-assisted models to solve business questions promptly & efficiently and attend to the evolving needs of the business when it comes to analytics.
Augmented Analytics solutions leave more space to the Data Scientist (being Expert or Citizen) to focus on the savviest part of the analytics’ project which is the insight and recommendations parts. This will in turn allow to add a significant value by automating the bottleneck jobs like data cleaning and preparation, especially that unstructured data environments like data lakes are increasingly adopted. Augmented analytics algorithms detect schemas and catalogues data and even recommend enrichment.
Are the human data science and machine learning skills going to be the same for the foreseeable future?
As clear as it can be, a paradigm shift in the analytics technologies field is happening and this implies that organisations need to take strategic moves to take the most out of it
It seems that a new ‘breed’ of data scientists is evolving to meet the needs of this era. This is what we call Citizen Data Scientists that are now gaining more and more notoriety and seem to be the perfect match for harnessing the power of augmented analytics.
But where do organisations have to look for Citizen Data Scientists?
Companies should consider upskilling existing employees or fresh graduates who have educational background in Physics, Mathematics, Finance, Computer Science, Economics… In other terms, anyone who is able and willing to perform quantitative work and who is aware of core business issues.
A combination of Citizen Data Scientists and Augmented analytics seems to be the winning combination, but in order to be successful, organisations need to prepare the adequate ecosystem, by raising data literacy and providing access to data use and most importantly to ‘speak’ data across the organisation.
by Oussama Mabrouki and Ramla Jarrar