A few days ago, I hosted a webinar on creating a data science portfolio that gets you noticed.
It was a fantastic session with the Pyladies Amsterdam community, and I’m excited to share some key takeaways about the step-by-step process to build a portfolio that can help you level up your transition to a career in data science. I’m sure these insights will equip you with the tools to build a portfolio that truly shines.
Or, if you prefer, you can watch the recording using the link below.
Step 1: Skills + Your desired industry.
To create a successful data portfolio, begin by identifying the specific skills you want to highlight based on your desired industry.
This initial step is crucial for clarifying the type of dataset you’ll need so your chosen project aligns with your career goals.
First, consider your target industry in Data Science: Are you interested in natural language processing for legal applications, machine learning for sales optimization, computer vision using satellite images, financial forecasting, or analyzing time-series data for wearable technology? The possibilities are endless!
Next, pre-select the types of datasets. Will you need text, numerical data, images, videos, maps, or a combination of these?
Basically, by defining your dataset requirements, you can focus your search and gather the most relevant data.
Step 2: Use your information advantage
For those transitioning careers, I’ve consistently seen positive results from clients who utilize their “information advantage.”
It means leveraging your knowledge and experience in the industry before switching careers.
You may not realize it, but you already know your field’s top questions and challenges, and that, my friends, is a great benefit!
Please don’t waste your years of experience; use them to your advantage!!
Take some time to write down the key questions you know are crucial to your current (or previous) industry. These questions will be invaluable for the next step.
Step 3: Search the dataset
Searching for a dataset when you already have a rough idea of what you’re looking for makes this step so much easier.
Use
- the questions you wrote down in step 2 and
- the type of data set you pre-defined in step 1
And look for data sets that align with those guidelines.
Download your dataset from places like Kaggle, Google data search and Data world.
Step 4: Data analysis
I always start with an initial exploration of the dataset.
Think about the questions you can ask and the information you can extract. Then, analyze the dataset, looking for the answers to each question.
Make sure to register your analysis in detail because that will be valuable in the next step.
Step 5: Storytelling and visualization
I love interactive plots for visualization that clearly show all the relevant information.
For me, one of the best things when it comes to visualization is that we have many options out there that are free and open-source.
Some of my preferred libraries are Bokeh, Matplotlib, Plotly, and Anywidget. I particularly like projects made by the community, such as Vega-Altair and Drawdata.
Step 6: Sharing your portfolio
The last step is to wrap up your projects and share your portfolio.
Make it easy to share. Basically, you only need a link. Keep it professional, but definitely add your personality.
Include an about me section with your contact information or links to social media profiles.
I suggest adding a list of your skills or grouping your projects by skills.
You can also add a few lines detailing your educational background and experience.
I hope these steps help you on your journey to create an amazing data science portfolio. I’m sure that with some effort, you can build a portfolio that truly showcases your skills and helps you land your dream job.
Lina Marieth xx
What’s New
Tool recommendation: Add Color Design to your creations
If you are working on your portfolio and want to add some color design, I found an amazing tool called Gradient Generator that will help you design a beautiful portfolio full of matching colors.
The tool itself is used to generate a gradual change in the color gradient from one color to another, essentially leaving the user with a result of many different in-between colors of the blend.
In addition to displaying the two component colors, as well as the various blends between them, the Gradient Generator tool also allows the user to choose the number of in-between colors displayed ranging from a single color to forty different in-between colors.
Isn’t that cool?
Quote of the week
Maybe the hardest part about taking a risk isn’t whether to take it, it’s when to take it. It’s never clear how much momentum is enough to justify leaving school. It’s never clear when it’s the right time to quit your job. Big decisions are rarely clear when you’re making them—they’re only clear looking back. The best you can do is take one careful step at a time.
― Alex Banayan, The Third Door.