Data Science is an evolving domain and over the last decade it has moved from a ‘good-to-have‘ to a ‘must-have‘ in running a customer-focused service or business. Like with any evolving field, its boundaries are getting pushed, skill sets are reimagining themselves, and notions about what a ‘Data Scientist’ is are constantly being challenged. Not surprisingly, businesses have re-organized themselves to make the best use of this game-changing asset.
With this we have seen the emergence of an Applied Data Scientist -a seemingly new job description. But is it? It has been there on the horizon for a long time, has gone through naming convention changes and continues to be one of the most powerful and widely required skill sets in the industry. We have known them as ‘Data Analysts’ in the past, now we may sometimes refer to them as ‘Data Scientists’ or in some organizations be specifically called ‘Applied Data Scientists’.
Who is an Applied Data Scientist
An Applied Data Scientist, simply put, is someone who studies the data (i.e. data science) with the aim of providing actionable solutions to business problems by applying theoretical conceptual frameworks and algorithms on the underlying data.
Someone who processes, analyzes, models and interprets data of any kind to drive meaningful insights and help solve business problems (and in most cases identify some more!).
Applied Data Scientists play with the underlying data, applying suitable technical/logical concepts leading to meaningful solutions to business problems .
An Applied Data Scientist is someone who:
Is skilled in ‘applying Data Science’
Being a good data scientist is the base of being a good applied data scientist. It goes without saying, ones does need to have a good understanding of the data one is dealing with. Also, this often means having a :
- Strong understanding of data science algorithms ranging from simple basic analysis to advance machine learning models. Its beneficial but not mandatory to have an in-depth understanding of the workings for all of these. Like with many other domains, Pareto’s Principle holds true in data science as well, and 20% of algorithms or approaches can be used in solving 80% of business problems.
- It is important to know these algorithms and have an awareness about their industry wide possible applications.
- Solid grasp on the platforms (or tools / software’s) needed to implement them. Understanding data science algorithms is of not much use if you don’t know how to use them on the underlying data. This is done via coding platforms that have machine learning / statistical packages and also via visualization tools which help us interpret and analyse.
- It is vital to know how to -read the data, process it, clean it, manipulate it, visualize it and model it.
Technical knowledge of approaches and the ability to implement them is a MUST HAVE
Has a firm grip on Domain knowledge
Domain knowledge or Industry knowledge is a key for Applied Data Scientists. During analytics projects, there often comes a time when one needs to steer the analysis in one direction vs. the other, that’s where industry know-how can help one apply business context/logic and steer the analysis. An applied data scientist has a sound understanding of his domain.
Domain knowledge goes a long way in taking decision calls during an analysis and can often be the difference between Good and Great.
Brings Insight to Action
If there is one thing that differentiates an Applied Data Scientist, it is the ability to make actionable recommendations to business problems after applying data science techniques on the underlying data. After playing with the data and using their domain knowledge on top of it to contextualize, an applied data scientist brings the analysis to life by making a story out of it. A story, which ends with recommendations based on the insights from the underlying data, which can be actioned upon by the stakeholders.
For example, creating a customer segmentation in retail is a piece of work that involves complex data science. However, in order to land this with business stakeholders, it helps to paint a picture of what type of customer’s these segments represent. Addressing questions like – what do they look like, what demographic they mostly belong to, how do they shop, what they shop etc. help the business to understand them, visualize them and bring the analysis to life (quite literally). The cherry on the cake is when data scientist’s can overlay their domain knowledge and describe strategies, which the business can implement, to activate these groups.
‘The What? So-What? and Now-What?’ framework is often used to bring the analysis to life and can help decide the key actions based on insight.
Also, Applied Data Scientists work very closely with clients (or business stakeholders). This is one reason they have a strong hold on domain knowledge and are able to apply business rationale to data science problems. Having strong relationships with business (or client) counterparts also mean that they are able to know what portions of their analysis can be actioned upon and how.
Developing and nurturing strong relationships with client counterparts is an essential ingredient for continuously delivering great client value.
Makes the complex, simple
Some Data Science algorithms can be very complex for a layperson to understand. The math involved simply can be very tricky to grasp and this is more true if one is not from a data science background.
Often the key stakeholders for data science projects are people who are decision makers, and they may/may not be from a data science background. This makes the job of a data scientist really difficult as it is very important to get stakeholder’s buy-in to ensure the analysis steers the decision-making process and not just ends up becoming a ‘good to have study’. This involves the ability to explain the complex science in simple layman terms to the key stakeholders and getting them to trust the analysis and hence, its findings. This ‘simple layman‘ portrayal of complex mathematical data frameworks/models can vary based on the stakeholder’s knowledge of data science or inclination towards understanding it.
Simplicity of a complex data science project ensures its wide reach and expanded usage.
Knows – Great vs. Good enough analysis
Applied Data Scientists sometimes work on projects which are short-term (2–3 days), sometimes long-term (multiple weeks) projects, and often work across multiple projects of different lengths. This broad range means that sometimes there are quick asks/requests from stakeholders where time is of the essence. In order to approach these, it is expected that the Applied Data Scientist has a well rounded knowledge of the existing products / bespoke solutions within the business; and is able to bank upon them to provide quick solutions in these scenarios. This obviously means interacting with other verticals (like Products team, Data Engineering teams, commercial teams etc.) on a regular basis and keeping a tab on the inter-connectivity of work across these.
It is almost always expected to not start from scratch and utilize the knowledge bank / tool set existing at your disposal
Also, if projects which are time sensitive mean taking a call on what is great vs. good enough? Knowing where to go deep into the analysis utilizing the best in class scientific algorithm vs. when to use a simpler framework and get the analysis done in time to aid the decision making.
When a business outcome is dependent on analysis, aiming for great solutions overtime is good and delivering good enough in time is great.
Maybe not NEW but it is Different
Applied Data Science profiles are certainly not new profiles in the Data Science world. However one thing that is different (.. and for the better), is that it is more precisely defined and better differentiated than a generic Data Scientist’s profile.
An Applied Data Science profile has a lot of benefits:
- The obvious one is learning – one gets to learn about Data Science, its applications, key tools as well as develop a solid domain knowledge overtime.
- Well defined roles means the people in the role are more clear around organization expectations, can set clear objectives and achieve them.
- Due to working closely with the clients, they are often able see decisions being made at the back of their analysis and can see the impact of their work readily.
- Working with multiple stakeholders across projects, this often means Applied Data Scientists have a strong network across multiple verticals.
- Applied Data Scientists can often carve out different career paths for themselves later on in their Careers based on the numerous skills they have picked up (and their wide network is obviously very helpful)
Applied Data Scientist vs. Research Data Scientist
Some organizations have a single ‘data scientist’ profile, while some have a segregated two profiles of data scientists – Applied Data Scientists and Research Data scientists (please note, that any one of these profiles can also be named just ‘data scientist’).
Are Applied Data Scientists different from Research Data Scientists?
This is a million dollar question. In my view, there is not much of a difference between the skill sets needed for these two profiles and the two can be inter-changeable when they want to be. However, due to the very nature of the ways of working or the organization’s expectations, there sometimes is a fine line between them.
But again, it must be emphasized that at the core there is NO reason for the two skillsets to be different.
Let’s looks around this ‘fine line‘ to understand it better:
Data Science Skills: There is no reason for the Data Science skillset across the two profiles to be different. The organization expects the Research Data Scientists to go deep into the problem (i.e. vertical depth is important), while the applied ones are expected to have a well rounded view of a wider variety of Data Science problems and solutions (i.e. horizontal coverage is more valuable)
- By the nature of their job, Applied Data Scientists might not know ALL the Data Science algorithms to the Nth degree whereas the research ones would more or less have a thorough understanding.
- Applied ones would be more skilled on broad applications of the 20% of concepts which can help solve 80% of Data Science problems in their domain.
- Both are expected to deal with complex advanced analytics, Machine Learning, Python etc.; however the Applied Data Scientists will always have the practical applicability in mind whereas the Research Data Scientists will go deeper into the problem.
Client Interaction Skills: Again, no particular reason for these to be different. Both profiles require a detailed understanding of business problems, the ability to turn them into Data Science problems and provide data backed solutions to them. However, due to nature of the job the Applied Data Scientists would have a greater proximity to stakeholders and hence would attain a wider/deeper view of business objectives giving them the ability to drive actionable outcomes from the analysis
In my view it is beneficial to have two separate job profiles and give the option to individuals to go deep or wide depending upon their aspirations. Obviously, having clearer expectations from employees is very helpful for the organization as well as the employees.
Summary
Interesting, Dynamic and Rewarding – these are words that come to my mind when I think about Applied Data Science in today’s world.
→ It seems to have a perfect balance of applicable technical expertise and business knowledge to do impactful work.
Connect, Learn & Grow ..
If you like this article and are interested in similar ones follow me on Medium, LinkedIn, connect with me 1:1, join my email list and (..if you already are not..) hop on to become a member of the Medium family to get access to thousands of helpful articles.****(I will get ~50% of your membership fees if you use the above link)
.. Keep learning and keep growing!
