The Data Analytics Project Workflow

Author

Karina Samsonova
October 31, 2025

You've learned SQL. You've mastered Excel. You can build dashboards.

But then you get your first real project at work, and suddenly you're staring at your screen thinking: "Where do I even start?"

Today, let's break down the actual workflow of a data analytics project.

Step 1: Understand the Business Question (Before You Touch Any Data)

This is where most analysts go wrong. They jump straight into the data.

Don't.

First, talk to your stakeholders:

  • What decision are they trying to make?

  • Why do they need this analysis now?

  • What will they do with the results?

Spend some time to gain clarity. It will save you hours later.

Step 2: Define Success / Definition of done

Ask: "What does a good answer look like?"

Is it a single number? A trend over time? A comparison between groups? A recommendation?

Get alignment on deliverables before you write a single query. What data are you expected to present and what is the format for deliverables (simple email, PowerPoint etc.)?

Step 3: Find and Assess Your Data

Now you can look at the data. Ask yourself:

  • Do we have the data needed to answer this question?

  • How clean is it?

  • What's the date range available?

  • Are there any obvious gaps or issues?

Sometimes you'll discover the data doesn't exist or it is archived and you have to request data recovery (my pain from this week). Better to know this on Day 1, not Day 5.

Step 4: Clean and Prepare

The unglamorous part. But necessary.

  • Standardize formats

  • Remove duplicates

  • Create any calculated fields you need

This usually takes 50-70% of your project time. Yes, really.

Step 5: Analyse (Finally!)

Now you can actually answer the business question.

Start simple:

  • Basic aggregations

  • Trends over time

  • Comparisons between groups

Then go deeper if needed.

But always tie it back to the original business question.

Step 6: Validate Your Results

Sanity check everything:

  • Do the numbers make sense?

  • Do they align with what the business expects?

  • If not, why not? (Sometimes unexpected results are the most valuable)

Run your analysis by a colleague if possible. Fresh eyes catch mistakes.

Step 7: Create Your Story

Data without context is just numbers.

Your deliverable should answer:

  • What did you find?

  • Why does it matter?

  • What should we do about it?

Choose the right format:

  • Dashboard for ongoing monitoring

  • PowerPoint for one-time strategic decisions

  • Written report for documentation

Remember: your slides are highlights, not a data dump.

Step 8: Present and Discuss

This is where your communication skills matter.

  • Start with the conclusion (executive summary). Sometimes this is what sparks a conversation and you don’t even move past the first slide

  • Show the key insights

  • Be ready to go deeper on what interests them

Step 9: Document and Handover

Future you will thank you.

Document:

  • Where you got the data

  • Any assumptions you made

  • Any cleaning tricky steps you made (removed duplicates, split data, filtered out specific dates)

  • How to refresh/update the analysis

  • Known limitations

If someone asks about this in 6 months, you should be able to recreate it. No need to be ‘War and Peace‘ length, it can be several bullet points.

The Real Workflow (Let's Be Honest)

In reality, it's never this linear:

  • You'll discover data issues in Step 5 and go back to Step 3

  • Stakeholders will change their mind halfway through

  • You'll realise you answered the wrong question and start over

  • The data will be messier than you thought

That's normal. That's the job.

Common Mistakes to Avoid

  1. Starting with the data before understanding the question

  2. Spending too much time making things perfect before getting feedback. Done is better than perfect

  3. Not documenting as you go (and hating yourself later)

  4. Forgetting to validate results with business logic

  5. Creating 50-slide presentations when 10 would do

Keep pushing 💪

Karina

Python tip

New way to merge dictionaries in Python 3.9+

dict1 = {'a': 1, 'b': 2}
dict2 = {'c': 3, 'd': 4}

# Old way
merged = {**dict1, **dict2}

# New way
merged = dict1 | dict2

print(merged)

Grab your freebies if you haven’t done already:

Data Playbook (CV template, Books on Data Analytics and Data Science, Examples of portfolio projects)

Need more help?

Just starting with Python? Wondering if programming is for you?

Master key data analysis tasks like cleaning, filtering, pivot and grouping data using Pandas, and learn how to present your insights visually with Matplotlib with ‘Data Analysis with Python’ masterclass.

Grab your Pandas CheatSheet here. Everything you need to know about Pandas - from file operations to visualisations in one place.

More from me: YouTube | TikTok | Instagram | Threads | LinkedIn

Data Analyst & Data Scientist