A Step-by-Step Guide to Transitioning your Career to Data Science – Part 2
How do you identify the technical skills a hiring manager is looking for? How do you build a data science project that draws the attention of a hiring manager?
This is a continuation of my previous blog post on the same topic. Make sure you have read that one before continuing here.
Step 5: Build your portfolio project
Your prospects have answered all your questions. Now, you have to analyze them and answer these two questions:
- “What business problems are my prospects working on?”
- “What are the main technical skills they posses?”
That’s it. Nothing more.
Coming back to my example, if most of my prospects say that they analyze web site data from google analytics and present these insights to the director of digital marketing. And they predominantly use SQL and Tableau to perform this analysis but rarely use Python or R. Adding to this, if they also mention that most of the Marketing Data Analysts learn R or Python on the job.
Then I know that I have to start learning SQL and Tableau. If I had started learning Python or R, in the first place, I would have wasted my time.
Your next step is to quickly build a portfolio project based on the answers to your two questions in this section. Don’t immediately sign up for a course to learn the technical skills required to build your project. Instead, you should learn enough technical skills to just complete your portfolio project. I learned this approach from Josh Kaufman, the author of the book "The First 20 Hours: How to Learn Anything . . . Fast!".
For my project, I am going to use Google Analytics Customer Revenue dataset from Kaggle. I then install Microsoft SQL server on my laptop and then upload my dataset to the SQL server. I am doing this because based on my answers I know that Marketing Data Analysts use SQL to retrieve data. So I want to demonstrate that I know SQL.
Your first step in your project is to understand the dataset and come up with 2 to 3 interesting business questions you can answer by analyzing it. You also need to know what columns (or features) in the dataset you need to answer these questions. This is where your domain knowledge comes into play.
Don’t try to complicate this project, your goal is to build a project that gets you the job. I call this portfolio project as “Minimum Viable Project”.
For my Marketing Data Analyst role, I must learn SQL and Tableau. So, first I learn the basics of SQL and Tableau. You shouldn’t spend more than 10 hours on learning the basics.
Once you learn the basics, your next step is to get in touch with a mentor who can help you to create a detailed plan on how to approach your project in a step by step manner. Choose a mentor who has good experience in data science. You can use services like Clarity or Mentor Cruise to find experienced data scientists. Schedule a call with your data science mentor and ask them what’s the best way to approach the problem. Explain the business questions you are trying to answer in detail.
After you receive a concrete plan on how to answer your business questions, start working on your project.
If you are stuck somewhere while working on your project or if you can’t find answers to your technical questions on forums like Stack Overflow, then again you can schedule a call with your mentor and get help from them. You can also find experts on CodeMentor to get technical help.
After you complete your project, once again review the project with your mentor and get their feedback.
While doing all this, you should also stay in touch with your prospects. Just send them an interesting article based on your previous conversation with them:
Saw this article in KDnuggets and it reminded me of what you said about Marketing Science! No response needed, just thought you might find it interesting.
Step 6: Get introductions to hiring mangers
Once you are done with your project, host it on Github. And write an in-depth blog post on how you approached the project. While writing your blog post, keep the technical details to minimum because you are writing this blog post to target business managers in the hiring team. A hiring team will consist of both technical and business people. The technical managers will go through your code on Github but the business managers are more interested in what business questions you are trying to answer using your analysis.
Email the link to your project and blog post to your prospects and ask for their feedback. If most of your prospects give a certain feedback, then implement that in your project. Now, your prospects will also say that they will refer you if a job opportunity comes up.
Thank them and but don’t send your resume. Instead, ask them to introduce you to someone in their team who has more experience. Here's a sample email that you can use:
Thank you for your feedback. I implemented your suggestions in my project.
In our chat last time, you mentioned [name of the person] would be the person to talk to about a potential Marketing Data Analyst position at [company name]. I'm definitely interested, and I'd love to chat with her/him. Would you be open to connecting me? I promise to be respectful of their time.
If so, I can send you a pre-formatted introductory email to make things really easy for you. If not, no worries -- thanks again for all your help!
Your goal is to get in touch with a senior person in their team. Someone who has the authority to hire you.
If your prospect agrees, you're in — expect a meeting or call with the hiring manger 80-90% of the time. Your prospects will most likely make the introduction because you are taking action based on their feedback. And you are considering them as a mentor. So it’s more of an mentor-mentee relationship.
With a proper introduction, a VIP will agree to speak with you almost every time.
At the actual meeting, treat it like an informal meeting— not a pitch, but an informal conversation: Ask questions, relate to your own experiences and oﬀer any ideas you may have. Also, talk about your project during the conversation but don't brag about it.
Most importantly, don't put them on the spot by asking for an interview yet, because you don’t know that they feel the same way. Just say, "Thank you so much for meeting with me. I'm going to take some time to think about all of this. Would it be all right if I followed up in a few days?" They will say yes. Then, send an email expressing your interest and ask THEM what they'd recommend as next steps. If they liked you, you're virtually guaranteed an interview — and you've already had coﬀee with the hiring manager. That's massive.
So, what’s the hiring manager thinking?
Again, put yourself in the hiring manager's shoes. They have big problems, but good help is hard to ﬁnd. If only they could ﬁnd someone who "gets it" and can hit the ground running. Enter you, the smart, discerning candidate who understands their world. It's okay if there are holes — you're clearly a quick learner, and you've done your homework (your project). They think, "Wow, he seems smart, normal, talented... we should at least bring him in to learn more."
Send a thank you to the hiring manger after the meeting. In the thank you email you should also provide a link to your project. Also, keep your prospect in the loop. Here's an email that you can send to your prospect:
Just wanted to let you know I met with [hiring manger] from [company name] the other day, and it went great! He invited me to apply for the Marketing Data Analyst position and I’m going to do so later this week.
No response needed; just keeping you in the loop. Thanks a ton -- I couldn’t have done it without you!
Step 7: Prepare your Resume and crack the interview
Now, it’s time to prepare your resume and cover letter. Your resume and cover letter should be focused on the job that you are applying to. Infuse the words of your prospects and hiring manager in your resume. That’s why it’s important to take notes during your meetings.
A resume only really needs these two main sections:
Here’s a great video by William Chen, data science manger at Quora, on how to prepare a data science resume.
All the work that you front-loaded pays oﬀ big time in the interview room, because you've:
- Performed deeper research
- Made better connections
- You have a project specific to their needs
80% of your work is done before you ever walk into the interview room. Now, it’s time to do the rest 20% of the work.
At this point, your prospective employer is curious about you. You put in an impressive application, and you were even introduced by a mutual contact. Now they want you to come in and meet the team, answer some questions, and just chat for a bit. Also, they want to get back to work! They want you to succeed — if you're right for the job. So prepare well for the interview.
Prepare a script for some of the most commonly asked interview questions like:
- "Tell me about yourself."
- "Why do you want to work here?"
- "Tell me about your previous work experience"
- “Talk about the project that you sent us”
Crack your interview and get your first dream job in data science.
I understand that this whole process is so much more challenging than mindlessly submitting resume after resume, but so much more rewarding in the long run.
Let me know if you have any questions in comments!
- A Step-by-Step Guide to Transitioning your Career to Data Science – Part 1
- Data Science Projects Employers Want To See: How To Show A Business Impact
- Projects to Include in a Data Science Portfolio