Get Free Widget

Welcome to your guide to getting started in Data Science! If you like what you read, scroll to the bottom or select 'sign up for the series' from the top menu to receive this guide as an eBook to read in your free time, as well as receive our series of free resources to help you transition to a Data Science career.

_Data Science Diagram

An Introduction to Data Science & the role of a Data Scientist

Recognised as “The Sexiest Job of the 21st Century” by Harvard Business Review, Data Science is one of the most valuable and multi-disciplinary skill sets, combining computer science, mathematics,statistics and domain expertise to predict trends, optimise performance, improve decision making and increase business competitiveness.

Demand for Data Scientists will soar 28% by 2020

IBM recently reported that demand for Data Scientists will soar 28% by 2020. Despite the popularity of the job, businesses across various industries are facing a big problem: there are not enough qualified data scientists out there.


The report notes that Data Science skills are one of the most challenging skills to recruit, potentially disrupting ongoing product development and strategic planning if positions are not filled.  In May 2018 alone, there were 3,696 active UK-based Data Scientist vacancies advertised on Glassdoor's website alone, signifying vast demand for the profession. So it’s not too late to make the transition...


Nowadays there is an assumption that you need a computer science degree or Ph.D. to pursue a career as a Data Scientist. We can tell you now, that isn't the case. Research into job advertisements show that the majority do not require computer science degrees.  Instead, the focus is on industry experience. As much as 78% of advertisements require that the applicant has relevant work experience providing evidence that they have the required skills to be successful in the role.


As companies embrace the need to become 'data-driven', we'll see career opportunities in Data Science becoming increasingly available. Especially as this is an  exciting and ever-evolving field, and Data Science skills are transferable across all sectors.

 

Insight

[When applying for Data Science-related roles] you should be able to talk about the different types of problems you’ve worked on. Then draw examples from your work experience that connect up with the missions of the business. For example, can one of the techniques you have worked with be applied to build a product useful for the customer of the company? This indicates you can adapt to different problems and put the models that you have used in a business context, which is one of the most valuable things we like to see.

 

- Phil Cowans, CTO, Nested

 

Screen Shot 2018-08-02 at 16.47.25

Skills You'll Need to Succeed as a Data Scientist

 

What does it take to get one of the 'sexiest jobs of the 21st century?'. We've scoured the web and asked internal experts what are considered crucial skills for candidates applying for Data Scientist and Analyst-type roles.  


According to Quora, the top five skills a Data Scientist should have include the following:

  1. Programming

You should have prior experience of computer programming, should you be considering a career in Data Science, or an analytical/statistical field. If you've no prior experience, it's never too late to learn!


Our Senior Data Scientist, Kevin Lamagnen, says:  

"Choose a language and stick to it; at least for the beginning so you can focus on learning the concepts. Then you can learn new languages. Whilst Python and R are popular within the general Data Science community, we recommend Python as it can be used for many other tasks such as building a website, automating tasks, and more".


  1. Teamwork

A Data Scientist cannot work in isolation. They're often deeply imbedded within the organisation interacting with different teams and client projects to tackle data challenges and business solutions.


Traits include:
  • Self Awareness: creating impactful solutions to complex problems requires constant iteration of your work and a constant seek of feedback to develop.

  • Knowledge sharing: because the field of Data Science is fairly new, documentation on the different frameworks, languages and libraries continuing to be developed and discussed. So as a practitioner, it’s great to be open to sharing your applied knowledge and experiences with others in the community.

  • Collaborative mindset: you should have the ability to quickly test and prototype code in a production environment, follow industry best practices, and use collaborative workflows to work efficiently in a team environment.

  1. Product intuition

This is another important skill to have that's tied to the Data Scientist/Analyst's ability to perform quantitative analysis on the system. Having a strong understanding of your product and the business processes through which you obtain your data from is essential in a number of situations:


  • To interpret the data properly
  • Evaluate the problem to generate hypothesis
  • Engineer features - this is where having domain experience really matters!
  • Execute and evaluate the model performance
  • Interpret the results and share your findings

  1. Communication

Data Scientists will need to be equipped with the skills to visualise and communicate the results of specific models, to explain and justify their ideas, methods and problem solving approach at a decision making level.

This skill manifests itself in various ways, including:
  1. Insights - you should be able to provide insight in such a way that individuals of varying technical levels would be able to understand. Being able to illustrate your findings in a relatable way is often referred to as 'storytelling'.

  2. Visualising findings for your audience - a good communicator will find ways, other than simply using words, to illustrate their point/findings, usually in a visual graph representation.

To conclude, It should be now clear that technical skills aren't primarily what form the ideal Data Scientist. Whilst Data Science is a multi-disciplined field,  soft skills are just as - if not more - important than technical skills.

 

If you can’t take abstract business issues and derive an analytical solution, explain complex technical issues in layman's terms or have good people/communication skills, these are core areas to focus developing. These are skills which can be improved over time with practice, there are also various free resources to help with improving communication, teamwork and commercial skills.

A data culture isn’t just about deploying technology alone, it’s about changing culture so that every organisation, every team and every individual is empowered to do great things because of the data at their fingertips.

What do Organisations Look for When Hiring Data Scientists?

We've included a job description for a recent advertisement for a Junior Data Scientist position we've recently recruited for to give you an idea of what organisations will be looking for [do bear in mind, the role of a Data Scientist can vary in different industries and organisation].


Junior Data Scientist (Python)

Cambridge Spark’s mission is to transform tech education for professionals with an intelligent tutoring system. Our proprietary system brings together research in software engineering, source code analysis, education pedagogy and machine learning. Cambridge Spark has recently won a government innovation grant to collaborate with Oxford University on recommender systems research applied for education.

We're looking for a junior data scientist experienced with solving Data Science problems, analyzing data and prototyping machine learning models with the Python Data Science stack (Pandas, Scikit-Learn, Numpy, etc..).


Responsibilities:

This is an exciting role as we are a small fast-paced team. You’ll have  a variety of challenges such as improving our own adaptive learning platform by prototyping new features, providing Data Science consulting to our clients, develop machine learning projects for our students and teach Data Science concepts both in our bootcamp and to our industry clients.  

 

Key skills:

  • Strong Python skills are a must
  • Experience with the Python Data Science stack (Pandas, Scikit-learn, Numpy, etc..)
  • Experience with Machine Learning
  • Experience with the whole Data Science cycle (from defining the problem to implementing a solution)
  • Experience with  software development tools (version control, unit test,  etc...)
  • Communication skills are a must. As we develop tools for education and run a bootcamp, it’s a plus if our engineers and data scientists can also teach various topics (Python, Web-Dev, Machine Learning, Databases, etc..)

Bonus skills

  • Experience writing production ready Python code
  • Experience in teaching
  • Some experience with web-development

 

How do I Get There?

1. University - MSC in Data Science

Due to the inclination of demand from industry for data science skills, universities have begun to offer a data science masters track in the UK. These can often last around 1-2, depending on whether undertaken full-time or part-time.

 

Pros
  1. Expertise - You'll often by taught by lecturers which are researching in the field they're teaching in.
  2. Funding - You'll be able to access a loan through student finance, if you're in the UK.

Cons
  1. Curriculum - Whilst there are links to relevant industry research, universities typically don't address the all of the skills and competencies industry are looking for and curriculums can become outdated very quickly.
  2. Cost - A masters can set you back around£5,000-£15,000+
  3. Network - unlike other routes, you'll have to network yourself through events and social media.
  4. Time commitment - Masters routes aren't the best route for working professionals, unless you consider a part-time route. Classes can often take place in the evenings, which isn't the most practical option for everyone.

 

2. Data Science Bootcamps

Bootcamps provide a more practical introduction to the field and teach a industry-relevant  curriculum - introducing them to the theoretical sides, whilst allowing students to apply and hone in on their skills.  


At Cambridge Spark, we not only work with industry partners and our expert tutors from leading universities to bring cutting-edge research to our curriculum. Upon completing core modules, we give students the chance to work with project partners on real industry problems for six weeks. This allows them to apply their learnings, network with other professionals, build a portfolio, and present their data stories to a wide range of diverse professionals, from industry and others in the bootcamp.  


Pros
  1. Curriculum - As the content you'll learn is created with industry partners [especially with our bootcamp], you'll be job-ready by the time you complete the bootcamp.
  2. Exposure - You'll be exposed to leaders currently working in the field through tutorials and networking events.
  3. Career support - Most bootcamps offer support in transitioning careers and finding a role upon completion.
  4. Portfolio building - Bootcamps will typically set projects for you to apply learnings, as well as build up a portfolio employers seek at interview stages to evidence your skills.
  5. Expertise - You'll often be taught by industry professionals with either practical experience in the fields they're teaching, or have links to academia, bringing research at the forefront of academia to the curriculum and teachings.
  6. Network - Bootcamps allow you to meet with diverse people from industry and in-class, allowing you to make friends, and network with professionals to increase your prospects of getting hired.
  7. Mentorship - You'll often be assigned an experienced mentor, helping you to bounce some ideas and understand where you'll need to develop.
  8. Instant feedback - Cambridge Spark are the only bootcamp provider to provide instant feedback on the quality of its students code submissions, as well as suggest tutorials to help you improve via our knowledge assessment Teaching Engine (K.A.T.E.)
Cons
  1. Time - when classes take place can vary, depending on providers. You may find your social life take a bit of a dip whilst you're undertaking the course - though it's a worthwhile dip if you're serious about a career in the field.



3. Self-study/online courses

 

Online courses are appearing on a variety of websites. These days, you're able to locate lots of specialised courses which offer a free introduction into various aspects of data science.

 

Pros
  1. Portfolio building - Online courses typically tend to be specialised, as opposed to comprehensive, giving you the flexibility to work on specific skillsets or areas and build up your portfolio to address a specific need.
  2. Fast learning curve - You're able to develop an understanding on various topics and put learnings into action through activities. Some are also developed with leading tech companies or experts in the fields, giving you a better understanding of the applications of learnings and industry context.
  3. Cost - Online and self-study options are often cheaper, or -in most cases - free.
Cons
  1. Support -  many free options don’t offer comprehensive support for students.
  2. Applications of learnings - You'll often not get the support you need to know you're on the right track, unless you opt for a pricier option.
  3. Guidance - Most online options won't offer the career guidance and check-ins other routes offer.
  4. Network - working in isolation will often mean you won't be able to develop the industry contacts you'll often build through other routes.
  5. Motivation - studying alone can work for some, though it can be demoralising not having the support network you would have with other options, especially when you get stuck.

Screenshot 2018-08-03 14.47.17

Tips for Getting Started

1. Follow relevant blogs

There are lots of sites that offer free newsletters, packed with news, tutorials and tips to help you prosper in your new career.

2. Network, Network, network

Start building industry contacts and don't be afraid to connect with other just because they have a fancy title. Apps like Shapr offer professional networking opportunities, in a Similar fashion to a popular dating app. LinkedIn tends to be a safe bet -many professionals will be happy to be approached for advice and mentoring.

3. Choose a language and stick to it

We recommend Python as it's widely adopted in industry.

4. Start practicing and communicating your results

Kaggle is a good way to do this to get started, this can enable you to learn from previous winners and practitioners to get you in the right mindset, before starting out on your own competitions.  You should consider starting a blog to get used to communicating your work and learning - a good platform to do so is Medium.

5. Seek constant feedback: use the contacts and groups you follow to ask for feedback

This will help you develop and see where you're going right and wrong.  Finding the right people to mentor you can make all the difference. Whilst there is no 'right path' to follow, there's certainly better ways of going about things.  

A well-seasoned Data Scientist can be a great contact to bounce ideas off, get feedback and ask for advice off of. It's also why we offer two mentors (one internal, one external) for students undertaking our project during our Applied Data Science Bootcamp.

Sign up for the series!

You'll receive the following direct to your inbox:

  • Your 'Getting Started in Data Science' eBook
  • Free resources
  • Datasets to help you practice your skills
  • Insight into Data Science Interviews
  • Examples of how Data Science is applied in industry

Please complete the following form to sign up: