GameCo

Know the Data. Master the Market.

Fast Facts for Busy Reviewers

Goal:
Help GameCo understand global sales patterns by region, genre, and publisher to support smarter marketing, partnerships, and platform investments.

Tools Used:
Excel • Tableau • Jupyter Notebooks • Python (Pandas, Scikit-learn, Seaborn)

Key Wins:

🎯 Action games lead in North America, Sports in Europe, RPGs in Japan
🧩 Puzzle game sales are down, likely due to shifting player trends
🕹️ Nintendo dominates global sales, but gaps exist for other publishers on newer platforms
📉 Sales peaked from 2005–2009, then dropped sharply post-2010 (likely due to the rise of digital)

Strategic Takeaways:

Tailor campaigns by region: Action (NA), Sports (EU), RPG (JP)
Partner with underrepresented publishers to fill market gaps
Adjust inventory and promotions based on platform trends and lifecycle stage

Explore the Code Behind the Insights

Slides

GitHub

Tableau

The Challenge

GameCo had been operating under the assumption that video game sales were pretty consistent across regions. But once I got into the data, it was obvious that wasn’t the case. Sales trends, genre preferences, and even publisher activity looked different depending on where you were in the world.

That raised some big questions:

Why are Action games huge in North America but not in Japan?
Why do Sports games dominate in Europe?
Are publishers missing chances by not adjusting to these differences?
And what does that sharp drop in sales after 2010 really mean for the future of physical games?

My goal was to help GameCo get past the "one-size-fits-all" mindset and see how customizing strategy by region could actually lead to smarter decisions and better results.

Problem

GameCo was using the same playbook for all markets, but the data showed clear differences in what works where. Without adjusting for regional trends, they risk wasting money and missing out on real growth opportunities.

Tools Used

To explore trends and structure the analysis, I used:

Excel
Tableau
Python (Pandas, Seaborn, Statsmodels)
Jupyter Notebook

Key Insights

North America loves Action games, Europe leans Sports, and Japan is all about RPGs.
Sales hit their peak between 2005 and 2009, then dropped off fast after 2010.
Nintendo is still the global sales leader, but EA and Ubisoft are falling behind on newer platforms.
Puzzle game sales are down, possibly due to changing player interests.

Business Suggestions

Market smarter by region: Action in NA, Sports in EU, RPGs in JP.
Build partnerships with publishers that aren’t yet strong on next-gen consoles.
Adjust inventory planning based on what sells best on each platform.
Keep an eye on digital trends to stay ahead of where the industry is headed.

Data Cleaning Highlights

Before diving into analysis, I cleaned and structured the GameCo dataset so it would be reliable for both modeling and visualization. Here’s what I did:

🕹️ Standardized genre, platform, and publisher names
Cleaned inconsistencies like "Sports " vs "Sports" and standardized spacing and casing to avoid duplicate groupings.

🕹️ Renamed columns for clarity
Converted messy Excel-style headers (like JP Sales or EU Sales) into jp_sales, eu_sales, na_sales, and global_sales for cleaner code.

🕹️ Dropped missing values and low-value rows
Removed any games that had no global sales or missing platform/publisher data. These would have added noise to trend analysis and modeling.

🕹️ Verified totals and region splits
Used Excel and Python to cross-check whether the sum of NA, EU, JP, and Other sales matched the Global_Sales column.

🕹️ Created a clean version for Tableau
Exported a polished .xlsx version with consistent region labels and no nulls so it would import cleanly for visuals.

🕹️ Prepared a modeling version
Removed string columns, scaled numeric values, and split features/targets for machine learning.

💡 What This Helped With

Built cleaner bar charts and line plots in Tableau
Made it possible to run accurate clustering and regression models
Prevented messy duplicates and wrong totals in pivot tables
Allowed easy filtering by region or genre in both Excel and Python

Snapshot of Key Cleaning Improvements

Analytical Build

Once the GameCo data was cleaned and prepped, I moved into the deeper analysis phase. The goal was to move beyond basic trends and start answering key questions about regional preferences, publisher performance, and market shifts.

To do that, I used a mix of descriptive analysis, visual exploration, and machine learning.

🕹️ Feature Engineering

To prepare the data for clustering, I used StandardScaler to normalize the numeric sales columns. This step ensured that na_sales, eu_sales, jp_sales, other_sales, and global_sales were all on the same scale before running KMeans.

Only numeric columns were included, and all string or categorical data was excluded to avoid skewing the model.

This preprocessing made it possible to apply unsupervised learning accurately across regional sales patterns.

🕹️ Clustering Analysis

To explore hidden patterns across platforms and genres, I used KMeans clustering. After scaling the data and testing different cluster counts, I found that grouping the games into three clusters helped reveal how genre and platform patterns varied regionally.

Cluster 0 leaned toward high NA and EU sales (Action/Sports-heavy)
Cluster 1 captured titles with high global balance (popular everywhere)
Cluster 2 focused on Japan-dominant genres like RPGs

Key Insight:
Games that performed well globally didn’t always perform equally across regions. This supported GameCo’s need to move away from one-size-fits-all marketing strategies.

🕹️ Elbow Method: Choosing the Right K

To figure out the best number of clusters for KMeans, I used the Elbow Method. I calculated the inertia score (within-cluster sum of squares) for different values of k from 1 to 10.

The sharp bend in the curve at k=3 to k=4 helped me choose the right cluster count to model how publishers group by sales patterns.

Key Insight:
Picking the right number of clusters is critical. Choosing too few hides important patterns. Too many overcomplicates the model. The elbow curve helped find that balance.

Key Visuals & What They Show

🕹️ Genre Sales by Region

This bar chart from Excel shows how much genre preferences change depending on where the games are sold.
Action games lead in North America, Sports are the top pick in Europe, and RPGs dominate in Japan.

What it shows:
Global averages don’t tell the full story. GameCo needs to align marketing with what each region actually prefers.

🕹️ Game Sales Over Time

This line chart from Tableau looks at total game sales across regions from 1980 to 2020. There’s a clear peak between 2005 and 2009, then a big drop-off after 2010 that shows up across the board.

What it shows:
GameCo has to be ready for these shifts. What worked during peak years won’t work forever. This trend helped highlight the need to plan for digital and platform changes.

🕹️ Top 5 Publishers by Global Sales

This pie chart from Tableau shows which publishers dominate global sales. Nintendo leads by a wide margin, followed by Electronic Arts and Sony Computer Entertainment.

What it shows:
GameCo has to be aware of who holds the power in the market. This visual helped pinpoint which competitors are shaping the industry and where GameCo could carve out a niche.

🕹️ Publisher Clusters by Region

This Python chart breaks down the average sales in each region for the publisher clusters I created using KMeans.
Some publishers do well across the board, while others are strong in just one or two regions.

What it shows:
Not every publisher has the same reach. This helped build a strategy that looks at who’s winning where and how GameCo can fill the gaps.

Main Insights (Recap)

🕹️ Same Game, Different Players

GameCo thought sales looked the same everywhere, but the data showed something else. North America loves Action, Europe sticks with Sports, and Japan is all about RPGs. A single strategy won't work across the board. What sells in one region might flop in another.

❯❯❯❯ It's not just what sells. It's where, and why.

🕹️ A Peak That Didn’t Last

Sales hit their high point between 2005 and 2009, thanks to consoles like the PS2 and Wii. After that, numbers dropped fast. That matches up with the rise of digital stores, mobile gaming, and subscriptions. This wasn't just a dip. It changed the industry.

❯❯❯❯ If you want to keep up, follow the players.

🕹️ Publisher Gaps = Open Doors

Nintendo leads in global sales, but EA and Ubisoft have fewer titles on newer platforms. Capcom is still focused on older systems. That leaves space for someone to step in. GameCo has a chance to fill those gaps where others aren't showing up.

❯❯❯❯ Don’t just watch the top players. Look at who’s missing.

🕹️ When Clusters Speak, You Listen

Basic numbers didn’t say much, so I used clustering to go deeper.
The Elbow Method helped me pick the right number of groups. KMeans brought the structure. It showed clear market behavior that wasn't obvious before.

❯❯❯❯ If the story is messy, sort it until it makes sense.

From Data to Decisions

This analysis focused on understanding why certain games and publishers performed better across regions, and where GameCo had room to grow. Rather than looking at sales totals alone, the goal was to uncover patterns in genre preference, platform shifts, and publisher strategy that could guide smarter decisions.

Business Decisions Informed by the Data

01. Adapt game strategy to regional preferences
The data showed clear differences in what players gravitate toward by region. Action titles dominated North America, Sports performed strongly in Europe, and RPGs led in Japan. A single global strategy missed these nuances.

❯❯❯❯ Games perform best when they’re designed and marketed with regional context in mind.

03. Look for opportunity where competitors aren’t active
Major publishers like EA and Ubisoft had fewer titles on newer platforms, while others remained focused on older systems. These gaps created space for GameCo to enter markets with less competition.

❯❯❯❯ Growth often comes from where others haven’t invested yet.

02. Use clustering to spot hidden market patterns
Clustering revealed groups of games and regions with similar performance behavior that weren’t obvious from raw sales numbers. This made it easier to identify untapped segments and avoid one-size-fits-all decisions.

❯❯❯❯ Strategy is stronger when it’s built around behavior, not averages.

04. Plan for long-term platform shifts
Sales peaked between 2005 and 2009, then declined as digital distribution, mobile gaming, and subscriptions reshaped the market. Focusing only on traditional consoles limited future growth.

❯❯❯❯ Companies that evolve with player behavior stay relevant longer.

What This Changes

This reframes GameCo’s challenge as a strategy and positioning problem, not just a sales problem. Strong performance depends on aligning games with regional demand, spotting gaps competitors overlook, and adapting early to platform shifts instead of reacting late.

Better decisions come from understanding where players are going, not just where they’ve been.