Which are the most important skills, needed by analysts or data scientists?
It’s an interesting question to ponder. It is tempting to fall back on technical basics, like statistical knowledge & software or coding skills. But most graduate data scientists or experienced analysts will be screened for those at recruitment.
My mind was prompted to consider these questions by two posts on LinkedIn. So, let me share those discussions and challenges with you. Then, as well as agreeing with the need for more UK research, I’ll share my suggestions.
Two analytics leaders lay down the “which skills” gauntlet
The first post to catch my eye was one shared by Martin Squires (Boots). It’s an IBM report published by Forbes, on data scientist skills shortages:
Jobs requiring machine learning skills are paying an average of $114,000. Data scientist jobs pay an average of $105,000 and advertised data engineering jobs pay an average of $117,000. 59% of all Data Science and Analytics job demand is in Finance and Insurance, Professional Services, and IT.
I agree with Martin that it focusses too much on technical skills & technical job titles. That said, the evidence of ever-growing demand is irrefutable. A survey I commissioned, with Worldwide Business Research, found similar challenges meeting demand. But that was focussed on other high-level challenges for data science leaders.
It is also interesting to see the salaries commanded by Data Engineers are higher than Data Scientists. Plus, technical skills like MapReduce top the ‘salary bumping’ list.
Like Martin, I want to see similar research for the UK. I’ll talk to a few of my contacts about the possibility of one of them conducting such a study. My own experience also confirms that it is normally non-technical skills that matter most.
So, I was encouraged to next see this post from Beau Walker (Liquid BioSciences). In it he poses the question of this blog post & shares his own views. The ensuing debate on LinkedIn shows no sign of stopping soon, as the relative merits of different skills are discussed:
What’s the most important skill for a data scientist in 2017? Expert-level Python or R? In-depth statistics or math? Data-wrangling? TensorFlow / Ke…
Although there are clearly two schools of thought in the ensuing debate, most appear to agree with Beau. Often, the most important skills for a data scientist in 2017 are still softer skills. As Beau puts it, the ability to interpret and communicate results.
My experience of the most important skills for data scientists in 2017
In two previous posts, I have made the case for Softer Skills and shared a 9 step model for effective analysis. Rather than repeat that material here, I’ll share my more recent reflections on the skills that matter most.
By way of background, especially for new readers, let me explain where I’m coming from. After a career spanning IT & then Underwriting, I spent 13 years creating & leading Customer Insight teams. Across Lloyds Banking Group & Scottish Widows, I created teams from scratch & developed others. These included data, analytics & statistical modelling teams (who would now be called Data Science).
After leaving corporate life, I created my own business and now help leaders in similar roles. Laughlin Consultancy provides training, consultancy, mentoring & other services. Working with data & analytics leaders to enable them to maximise the value of their teams.
So, my observations on the skills needed span both my 13 years as a practitioner & now 3 more years as a supplier. My experience across all those years has actually remained remarkably consistent. Despite all the advanced developments in Data Science, the skills that are still most needed in such teams are all too familiar.
Answering Beau’s original question, from all the different skills I teach, I’d select this top 3:
Most important skills: (1) Communication – inc. Data Visualisation
After training analysts in a full range of softer skills, the most in demand skills to develop further are communication ones. That also rings true with my prior experience. The analysts (or data scientists) who made the most difference were the best communicators.
This doesn’t have to be limited to extroverts either. You might expect more confident presentation skills from extroverts in your team. But introverts can also have strong skills in analytical thinking & visualisation of data. I’ve known plenty of more introverted analysts, who do such a good job of summarising key points & displaying data – that it almost speaks for itself.
Under this heading of broad communication skills, I would emphasise two elements in particular. General communications skills training is still of value. But two skills that are essential to a data scientist’s toolbox are storytelling & data visualisation. I’ve shared before about the importance of developing a big picture & recommended a book on storytelling in business.
Data visualisation skills can transform how easily non-technical audiences can digest analytics results. So, I’d recommend taking a look at the blogs, tweeters & books this blog has already recommended on that topic. Beyond those, I do recommend investing in face to face training for this skillset. One option for doing so is this workshop, part of DataIQ’s excellent Leaders package:
In this workshop, analytics practitioners & leaders will have opportunity to learn from data visualisation theory & the practical experience of a senior data insight leader. This practical workshop will explain both the importance of data visualisation and how to implement principles in practice.
Most important skills: (2) Stakeholder Management – inc. Influencing
Almost as often requested, are improved skills in managing difficult stakeholders. Everyone has some. Every analytics leader I have worked with, at some point highlights them. Normally, they are senior leaders who appear to not understand or care about data analytics. Often apathy, rather than challenge, is the bigger problem.
Stakeholder management may sound too generalist to apply to technical roles like data scientists. But I find technical experts need these skills more than most. Theirs is the challenge of making the complex simple & also needing to influence others to act on their findings. Most data scientists come to work to make a difference. Although the data analysis & modelling work may be under their control, they can hardly ever act on what they find in isolation. Others will own the marketing or customer service systems needed to execute new targeting or experiences.
When training analysts in Stakeholder Management, it becomes plain that many of those skills apply to them. Mapping your stakeholders, understanding their style, developing your voice & planning your tactics. All those skills may sound more like sales training. With Data Science still being new to most businesses, there is indeed a sales job to be done.
This topic is so important for technical leaders & analysts, it is going to be our theme for October. But in the meantime, I’d recommend reading previous blog posts on influencing C-suites and PR for your team. Once again, I also recommend face to face training on this topic. The opportunities to role play, or talk about your challenges with other delegates, can make a big difference. DataIQ Leaders are also providing a workshop on this next month:
The day includes plenty of interactive exercises. There is opportunity to apply what you are learning to your own business context & challenges. Working with others leaders, you can practice new skills and learn more about your own leadership development.
Most important skills: (3) Questioning – getting to business need
Looking at the feedback from past training courses, on Softer Skills for Analysts, the most popular exercise is on questioning. I learnt about Socratic Questioning during my qualification as a leadership coach. On that course, I explain what it is and offer some example questions to use. Then delegates get to have the fun of role play.
Not only can this sometimes be hilarious to watch, practicing questioning skills is often cited as one of the most helpful parts. Talking to analysts later, they will tell be they are still using questioning cheat sheets. Some teams have event decided to make practising of questioning skills a regular part of team meetings.
So, based on my client-side experience of influencing senior banking & insurance leaders, I recommend this as the last of my top 3. Honing your craft in asking incisive questions, that get to the real business need. That skills will add value to any analytical role.
You might find it helpful to take a look at our previous post on the benefits of Socratic Questioning. I don’t know of another training course that trains analysts in these skills. So, although I try and avoid pushing my services on this blog, I can only recommend my course:
From the softer skills that analysts need to drive action on insights, to specialist topics like Conduct Risk, Laughlin Consultancy can provide the training your team needs. More and more businesses are realising that you can’t just hire all the skills & knowledge needed. Why not invest in training your team?
Data Science leaders, which are the most important skills to you?
Did that help? Do you agree?
If you are leading a Data Science team, I’d love to hear your perspective. Whether you’d prefer to comment below or on social media, please share the skills you have found matter most. With Data Science still being a relatively new discipline in businesses, we are blazing a trail. Let’s help leave a clear path for those who follow.
Enjoy developing yourselves and your teams.