How do you break into Data Science? This blog is part one of a series on FAQ’s we receive from professionals looking to upskill in Data Science and Machine Learning to achieve just this.
FAQ #1: how do I break into the data science profession when I have no prior experience in the field?
As companies embrace the need to become “data-driven,” career opportunities in data science are increasingly available. It is an exciting and ever-evolving field, and data science skills are transferable across a range of sectors — so how can you get involved?
The elements of Data Science
- Domain knowledge to define the business problem
- Programming skills to apply the most appropriate statistical / machine learning techniques to the problem
- Technical knowledge and understanding to rigorously evaluate the solution and generate better outcomes
- Business acumen to communicate your findings and explain your problem solving approach at a decision making level
To acquire the skills needed for the job you need to develop theoretical understanding and get hands-on practice. Here are some steps you can take to prepare:
Being curious and doing further research into the field is important. You’ll need to dedicate time to go through tutorials, read books, even take online courses. It’s time consuming, but this routine will help you gain a strong understanding of how to use each technique and how you can apply them in the real world.
Competing in Kaggle is a great way to practice your skills, and you can check out code shared by other participants to learn and improve. The Titanic dataset is a good place to start, then move onto more advanced problems. However, keep in mind that Kaggle competitions are limited to “beat the benchmark” kind of tasks, whereas a data scientist spends more time understand the data, cleaning it and building maintainable models and code.
So the next step is to pick a dataset that you care about and work on a more realistic project. Define a project goal, then spend time analysing the data in depth and explain the assumptions in a notebook. Some open data sets can be found at: https://github.com/awesomedata/awesome-public-datasets
Focus on your code quality. Build good code that processes the data, allows to simply train your model and retrain it if the data changes, predicts on new data and is maintainable and readable by other engineers. This will make the difference when presented to employers.
Post your code to your Github to build your portfolio. You can also start a data-related blog to explain your work and practice writing about your methodology. Then when you’re more confident you can contribute to open source projects such as sklearn, pandas, etc.. to give back to the data science community.
At our Applied Data Science Bootcamps, we give you the opportunity to go through all these steps. Not only will you work through a series of intensive modules to cover all the theory and complete practical projects every two weeks. When our 10 raining modules end, you will begin working on an industry-capstone project. This is your chance to tackle an end-to-end data science problem with a company. Then after 6 weeks, we reunite for a project presentation showcase to evaluate your solution and give feedback.
Interested in learning more? If you need advice or have any specific questions, get in touch with our team at: firstname.lastname@example.org.