It's not the first time that someone reaches out and asks: "I've never done a data project before—how do I actually begin?"
I remember that blank feeling. No portfolio. No idea where to start. Just the wanted to get going.
Here's what I tell every beginner who asks me this:
Step 1: Pick a dataset that feels interesting. Don't overthink it. Something simple—maybe sales data, or your city's bike-sharing records. Kaggle, your gov data, and Google datasets are all great places to look.
Step 2: Define a single, clear objective. This is where most people freeze up. It can be as basic as: "Is there a monthly trend?" or "Which product sells best?" If you have a question you're curious about, that's your objective.
Step 3: Get familiar with your data. Open it up. Glance at the columns and values. Are there missing values? Does everything make sense? Write down your first impressions.
Step 4: Clean the data just enough to analyze it. Remove obvious errors, fill in or drop missing values, and standardize your columns. All you want at this stage is clarity.
Step 5: Run simple analyses — averages, totals, min and max, basic counts. Plot a few charts (bar, line, pie). Look for anything that stands out. This is the fun part.
Step 6: Summarize your findings in clear, plain words. "Product A sold the most in June." "More people use bikes on weekends." Keep it simple, honest and data-backed.
That's it. No fancy algorithms, no big words. You've just finished your first real project.
I'm guessing it should take you no more than 30 minutes. If it does, you're making things way more complex than what they have to be.
You don't need permission to start. Just pick something and go.
Share this with anyone who's struggling with data projects.
The Ultimate Resume Builder: If you're spending more than 2 minutes fixing your resume to fit specific job descriptions: mochen.info/lp-ultimate-resume-builder
Mo Chen
It's not the first time that someone reaches out and asks: "I've never done a data project before—how do I actually begin?"
I remember that blank feeling. No portfolio. No idea where to start. Just the wanted to get going.
Here's what I tell every beginner who asks me this:
Step 1: Pick a dataset that feels interesting. Don't overthink it. Something simple—maybe sales data, or your city's bike-sharing records. Kaggle, your gov data, and Google datasets are all great places to look.
Step 2: Define a single, clear objective. This is where most people freeze up. It can be as basic as: "Is there a monthly trend?" or "Which product sells best?" If you have a question you're curious about, that's your objective.
Step 3: Get familiar with your data. Open it up. Glance at the columns and values. Are there missing values? Does everything make sense? Write down your first impressions.
Step 4: Clean the data just enough to analyze it. Remove obvious errors, fill in or drop missing values, and standardize your columns. All you want at this stage is clarity.
Step 5: Run simple analyses — averages, totals, min and max, basic counts. Plot a few charts (bar, line, pie). Look for anything that stands out. This is the fun part.
Step 6: Summarize your findings in clear, plain words. "Product A sold the most in June." "More people use bikes on weekends." Keep it simple, honest and data-backed.
That's it. No fancy algorithms, no big words. You've just finished your first real project.
I'm guessing it should take you no more than 30 minutes. If it does, you're making things way more complex than what they have to be.
You don't need permission to start. Just pick something and go.
Share this with anyone who's struggling with data projects.
The Ultimate Data Portfolio: If you're looking for the best data portfolio for job seekers:
mochen.info/lp-ultimate-data-portfolio
The Ultimate Resume Builder: If you're spending more than 2 minutes fixing your resume to fit specific job descriptions:
mochen.info/lp-ultimate-resume-builder
3 days ago | [YT] | 250