7 Machine Learning Projects to Build Your Skills

Written by Coursera Staff • Updated on

Machine learning projects are a great way to practice your skills and develop your portfolio. Test yourself and prepare for a future career as a machine learning expert with these engaging projects.

[Featured Image] A machine learning student works on a machine learning project on their laptop in a library at a wooden table. They're wearing headphones and there is a stack of books next to their computer.

So, you’ve been developing your machine-learning skills, diving into the finer points of data points, and practising programming languages. What’s more, you know what a machine learning model is and want to get your hands dirty actually making one rather than just reading about it. 

Machine learning (ML) projects allow you to practice the skills you’ve developed while giving you something to showcase in your portfolio. As a result, they help you better understand data science and machine learning and demonstrate to potential employers what you can do when given the chance. 

To help you get started, here are seven machine learning project ideas suitable for beginners and more advanced ML students. 

1. Identify irises

Irises influenced the design of the French fleur-de-lis, are commonly used in the Japanese art of flower arrangement known as Ikebana, and underlie the floral scents of the 'essence of violet' perfume [1]. They’re also the subject of this well-known machine learning project, in which you must create an ML model capable of sorting irises based on five factors into one of three classes: Iris Setosa, Iris Versicolour, and Iris Virginica

To help you get started, the data set below includes 50 instances of each of the three iris classes for 150 cases. While one of the classes is linearly separable, the other two are not. Your task is to make a model capable of classifying each iris instance into the appropriate class based on four attributes: sepal length, sepal width, petal length, and petal width. 

UCI data set: UCI Machine Learning Repository Iris Data Set

2. Forecast sales

How will the changing seasons, shifting demographics, or government regulations impact a business’s future sales? 

Questions like this undergird the standard business practice of sales forecasting, in which a business estimates the number of products or services it will sell in the future based on relevant historical data. Unsurprisingly, companies have increasingly turned to machine learning to build models capable of forecasting sales with greater and greater accuracy than the less technologically advanced approaches of the past. 

In this machine learning project, you will gain experience with sales forecasting using real-world sales data provided by Walmart. Your task is to predict the department-wide sales for 45 Walmart stores in different regions while considering important seasonal markdown periods such as Thanksgiving and Christmas. 

Kaggle data set: Walmart Recruiting—Store Sales Forecasting

3. Predict stock prices 

A typical piece of investing advice is that the key to beating the market is to buy stocks at their lowest price and sell them at their highest. In other words, buy low, sell high. But how do you know when a stock is at a low point and when it’s reached its peak? 

While there is no foolproof way to answer this question, one approach is to develop a machine-learning model that can try to predict stock price fluctuations using historical data. That’s precisely what you will try to do in this machine-learning project. 

The data set below includes high-quality data for US-based stocks and exchange-traded funds (ETFs) on the NASDAQ, NYSE, and NYSE MKT. How might you try to crack the ever-elusive question of predicting stock prices with machine learning? 

Kaggle data set: Huge Stock Market Data Set

4. Design a recommendation system

We’ve all been there: you’re on a streaming platform with a seemingly endless collection of videos and unsure what to watch. Do you try that anime series set in the not-so-distant future or that cheesy romantic comedy clearly from the early aughts? Or, should you finally get to that atmospheric noir from the 1940s? 

Online platforms are aware of the decision fatigue that can result from an overwhelming number of options, so many employ complex machine learning models to tailor user recommendations. Recommendation systems underlie many of the most popular services today—from Google to Netflix to Xbox’s Game Pass service. 

In this project, you’ll create your recommendation system using data collected from the movie-recommendation service MovieLens. Created by 138,493 users, the Movielens data set includes over 20 million ratings and 460,000+ tags for 27,278 movies. See what you can do with this important data. 

Kaggle data set: MovieLens 20M Dataset

5. Predict house prices 

Buying a home is often one of the most important—and expensive—milestones in a person’s life. As a result, the real estate and housing markets are some of the most significant within the global economy. 

While financial value certainly isn’t everything when purchasing a home, many people want to know if a specific home will be a good investment in the long term. For example, how much might your house be worth if you sell it today or after you renovate it? 

With so much publicly available real estate data, predicting housing prices is a natural fit for a machine learning project. In this self-guided lab by Google Cloud Training, you will learn how to use machine learning to predict housing prices by building an end-to-end solution using Tensorflow 1.x and AI Platform. You’ll also learn how to leverage the cloud for distributed training and online prediction, in general, so that you can use your skills on future projects. 

6. Identify damaged car parts 

During the COVID-19 pandemic, supply chains were stopped globally as countries and workplaces shut down to prevent the spread of the virus. As a result, the automotive industry struggled to manufacture new cars.

As a potential car buyer during that period, you’d likely be concerned about the condition of a possible car purchase as you scrolled through used car listings. Wouldn’t it be great if you could use machine learning to identify the damage to different car parts and determine if the purchase would be worth the investment? 

In this interactive project by Google Cloud Training, you will do just that as you use machine learning vision to identify damaged car parts. Designed for intermediate machine learning practitioners, this quick project will walk you through uploading a data set to cloud storage, inspecting uploaded images to ensure no errors, training an ML model, and evaluating your model for accuracy. 

7. Identify emotions

As painters, sculptors, and actors have known for millennia, the face is a wellspring of emotion. While actors in traditional Japanese Noh theatre use light and shadow to convey smiles and frowns on otherwise unchanging masks, the ancient sculptor who created the famous statue Laocoon and his Sons used contorted expressions on his subjects’ faces to convey their suffering as snakes attacked them. 

Then, the face and its expressions are another data source—often intuitively understood by many humans but not so by machines. Nonetheless, the key points on faces that alter as expressions change allow machine learning models to identify at least some emotions.

In this Coursera project, you will use Artificial Intelligence (AI) to predict emotions based on different facial expressions. In this three-hour guided project, you will build and train a deep learning model based on Convolutional Neural Network and Residual blocks using Keras with Tensorflow 2.0 as a backend.

Reinforce your learning (and gain new skills)

Machine learning is a growing field with a wide range of applications. Whether you are just starting or are already well acquainted with the field, Coursera has a course for you. 

Stanford University and Deeplearning.AI’s joint Machine Learning Specialisation can help you master fundamental AI concepts and develop practical machine learning skills in a beginner-friendly, three-course program by AI visionary Andrew Ng. Meanwhile, DeepLearning.AI’s Deep Learning Specialisation teaches intermediate-level course takers how to build neural networks, CNNs, and RNNs. 

Article sources

  1. Britannica. “Iris, https://www.britannica.com/plant/Iris-plant-genus.” Accessed March 25, 2024.

Keep reading

Updated on
Written by:

Editorial Team

Coursera’s editorial team is comprised of highly experienced professional editors, writers, and fact...

This content has been made available for informational purposes only. Learners are advised to conduct additional research to ensure that courses and other credentials pursued meet their personal, professional, and financial goals.