There’s so much content on the internet outlining the path to become a data scientist. Most of them follow a similar pattern: Here are five skills to learn to become a data scientist, learn python from here, do these five projects to get a data scientist job, and so on.
There’s a gamut of MOOCs and data analytics certifications and short-duration courses that claim to teach data science from scratch: Coursera, edX, Simplilearn, and more that claim to prepare aspirants to learn data science in a matter of few months and get them job-ready. Not to mention, a few programs even prescribe learning discrete mathematics, statistics from the ground up, and more. Some can definitely benefit from these; however, these aren’t effective ways to learn data science, especially if you’re trying to get a data scientist jobs.
Sure, you can follow instructions in these programs; like downloading the iris data set for building a classification model, learning tensor flow to build a deep learning model; before you give up, or go compete on Kaggle. These recommendations are good, but these are ineffective approaches. Many of these approaches incorporate ‘learning by doing’, by working on projects.
It’s understandable that we have the habit of following a structured pattern of learning ingrained in us. But when it comes to getting a job in data science, a better approach is to build a public portfolio that can you can show off. The following approach recommends
Here’s how you can start:
- Find a topic you’re passionate about
The premise behind the evolution of data science is: solve problems using data. So choose an area you’re passionate about – pets, weather, wildlife, economy, business, public policy, or anything. The idea is to find an area that will keep you motivated to work on the project. Once you have finalized an area, find a way to gather data. Perhaps you’re working for an NGO where you can get hold of data or find a place where you can easily gather data.
The idea is to answer interesting questions. For instance, which is the busiest hour in the city? Or where can you find the cheapest beer? Keep a time frame of a week and aim to build something in a week or so. Working on a project that you are passionate about will keep you motivated to build something fast. If the area you have chosen doesn’t interest you, repeat the first step.
- Send a tweet
By making your idea public, you put yourself in a position to complete a task. This method is used by Amazon. The company writes a press release first and then gets on with work. This is also an essential part of the lean start-up philosophy. A similar approach that you can follow is to write a tweet.
An example would be: “Hats off to the state government, the poverty in our state reduced much faster than the rest of the country. I took the data from the public office and compared it with the rest of the states in the country.”
Alternatively, you can document your idea and work toward realizing it. Discipline would be essential to complete the project.
- Do the work
Next, dig up data and employ all your skills. Collect, clean, and analyze data. Find if there are any inconsistences with the data. Analyze data and see if there are any missing values. If there are, can you find other data sets that can complement your work better? Check for outliers. Get feedback on your progress from Kaggle and experienced data scientists on LinkedIn. This work should be exciting and fun. If doing all this work sounds ‘meh’, it’s time to reconsider that you want to pursue a career in data science.
- Communicate
Data visualization is an essential part of data scientist’s work. So the next step is to communicate your findings. If you have a prototype, a video of the model would be nice. Additionally, keep your code, additional findings, and corroborating evidence just a link away. In this process, you will learn various data visualization tools.
Congratulations now you have a project that you can show off. Repeat these steps and work on more interesting projects of your choice. In the process, you will develop various skills and build competency.
- Get a data science certification
Though not mandatory, getting a data analytics certification will work in your favor too, especially vendor-neutral certifications. In addition to projects, these certifications will prove your competency as a data science professional. Your projects obviously show more, but certifications will accentuate that value.
Ready to work on projects? Get going!
