Artificial Intelligence
Sep 7, 2021

6 AI Projects To Make Your Portfolio Stand Out

Make your AI portfolio standout with these projects.

Shanif Dhanani

Data science is becoming hyper-competitive. One way to stand out is with a portfolio of valuable projects —not just analyzing the Titanic and Iris species.

Here, we’ll cover six projects to boost your portfolio:

1. Churn Analysis

Churn is when a customer quits a service. In other words: It’s when companies lose money.

When Blockbuster’s customers left for competitors like Netflix, they churned. Churn analysis could have uncovered the why and potentially saved the once-dominant media giant.

The goal of churn analysis is to effectively pinpoint the causes of churn and fight it. By directly improving the bottom-line, churn analysis is a powerful AI use-case.

The first step to churn analysis is creating a dataset (or simply finding the relevant data). Kaggle offers many churn datasets — a popular one is the Telco Customer Churn dataset, though this is synthetic and not as unique.

Less-explored datasets include predicting churn for bank customers, newspaper churn, insurance churn, AudioBook app churn, and tech company churn.

For any of these datasets, you can easily download them as a CSV file, and upload them to Apteo to find insights (after making an account).

2. Customer Segmentation

Customer segmentation lets you uncover the characteristics of different customers, enabling better targeting.

It’s a way to “organize” your customers, and understand the habits of these groups.

Photo by Craig Whitehead on Unsplash

Less-explored Kaggle datasets include wine customer segmentation and US supermarket segmentation.

3. Driving E-Commerce Sales

E-Commerce analysis is a great AI use-case, given the wealth of data available. You can identify browsing behavior that leads to increased sales, segment profitable users, identify pages likely to convert, and more.

Relevant Kaggle datasets include outdoor apparel data, cosmetics store data, Retailrocket data, Flipkart data, and more.

Companies from Clorox to PepsiCo use AI for data-driven e-commerce — it’s that powerful!

4. People Analytics

People analytics is used to analyze and reduce employee turnover, spot opportunities for improvement, and more.

As LinkedIn reports, “for each employee lost, the cost to the company could be 50%–250% of his/her annual salary.” While that’s a big range, there’s no doubt that employee attrition is extremely expensive.

Intuitively, we can guess that things like overtime, low salary, poor performance, and living far from the office might impact attrition.

With people analytics, we can plug-in the data to quantify the true impact of these attributes, and more, on attrition. With the insights we get, we can take steps to reduce attrition, whether it’s reducing over-time, increasing salaries, changing department structures, or altering company policy.

Relevant Kaggle datasets include the IBM attrition dataset, a recent HackerEarth attrition dataset, and an Oakland PD attrition dataset.

5. Lead Scoring

Lead scoring helps teams close deals and build relationships by prioritizing customers likely to convert.

Whether you have 130 or 130,000 leads, you need to put your efforts where they matter.

Relevant Kaggle datasets include the marketing funnel by Olist dataset and an education leads dataset.

6. Fraud Detection

Fraud detection is becoming increasingly important, as fraudulent transactions are only increasing, costing companies billions.

Combating fraud is an age-old problem. Back in 1992, Sprint was using real-time fraud detection built by Symbolics, an AI company. Today, companies like Visa, PayPal, and FirstAm deploy data-driven fraud detection.

Relevant Kaggle datasets include credit card fraud detection and synthetic financial datasets for fraud detection.


Photo by Sebastian Herrmann on Unsplash

About the author
Shanif Dhanani
Co-Founder and CEO, Apteo