Data scientists are numbers people. They have a deep understanding of statistics and algorithms, programming and hacking, and communication skills. Data science is about applying these three skill sets in a disciplined and systematic manner, with the goal of improving an aspect of the business. That’s the data science process.
In order to stay abreast of industry trends, data scientists often turn to case studies. Reviewing these is a helpful way for both aspiring and working data scientists to challenge themselves and learn more about a particular field, a different way of thinking, or ways to better their own company based on similar experiences.
If you’re not familiar with case studies, they’ve been described as “an intensive, systematic investigation of a single individual, group, community or some other unit in which the researcher examines in-depth data relating to several variables.”
Data science is used by pretty much every industry out there. Insurance claims analysts can use data science to identify fraudulent behavior, e-commerce data scientists can build personalized experiences for their customers, music streaming companies can use it to create different genres of playlists—the possibilities are endless.
Allow us to share a few of our favorite data science case studies with you so you can see first hand how companies across a variety of industries leveraged big data to drive productivity, profits, and more.
6 case studies in Data Science
1. Gramener and Microsoft AI for Earth Help Nisqually River Foundation Augment Fish Identification by 73 Percent Accuracy Through Deep Learning AI Models
Source | Gramener
The Nisqually River Foundation is a Washington-based nature conservation organization. They sought to implement a watershed stewardship plan, but first needed to measure and monitor the fish species present in the Nisqually River. To do this, they installed a video camera and infrared sensors in the water. The camera was triggered to record 30 seconds of video when any fish appeared. Later, the videos were reviewed and the fish were manually identified. This was a time-consuming, inefficient process, so when the organization decided to do this a second time for the salmon species, they turned to Gramener for an automated, tech-driven solution.
Gramener, a data visualization and predictive analytics company, implemented a web-based artificial intelligence (AI) program. It is projected to deliver the Nisqually River Foundation savings of up to 80 percent.
2. How We Scaled Data Science To All Sides of Airbnb Over 5 Years of Hypergrowth
One of the startup world’s profitable unicorns, Airbnb was one of the few companies that included a data scientist within its initial team so it could evolve as quickly as possible.
The company experienced 43,000-percent growth in just five years, so the strategy clearly worked. The goal of this case study is to share some of the high-level issues Airbnb encountered as it grew, and details about how its data science team solved them. To do so, the author (Airbnb’s first data scientist) breaks things down into three areas:
- How Airbnb characterizes data science
- How data science is involved in decision-making at Airbnb
- How Airbnb has scaled its data science efforts across all aspects of the company
Airbnb says that “we’re at a point where our infrastructure is stable, our tools are sophisticated, and our warehouse is clean and reliable. We’re ready to take on exciting new problems.”
3. Spotify’s “This Is” Playlists: The Ultimate Song Analysis For 50 Mainstream Artists
If you’re a music lover, you’ve probably used Spotify at least once. If you’re a regular user, you’ve likely taken note of their personalized playlists and been impressed at how well the songs catered to your music preferences. But have you ever thought about how Spotify categorizes their music? You can thank their data science teams for that.
The goal of the “This Is” case study is to analyze the music of various Spotify artists, segment the styles, and categorize them into by loudness, danceability, energy, and more. To start, a data scientist looked at Spotify’s API, which collects and provides data from Spotify’s music catalog. Once the data researcher accessed the data from Spotify’s API, he:
- Processed the data to extract audio features for each artist
- Visualized the data using D3.js.
- Applied k-means clustering to separate the artists into different groups
- Analyzed each feature for all the artists
Want a sneak peek at the results? James Arthur and Post Malone are in the same cluster, Kendrick Lamar is the “fastest” artist, and Marshmello beat Martin Garrix in the energy category.
4. A Leading Online Travel Agency Increases Revenues by 16 Percent with Actionable Analytics
One of the largest online travel agencies in the world generated the majority of its revenue through its website and directed most of its resources there, but its clients were still using offline channels such as faxes and phone calls to ask questions. The agency brought in WNS, a travel-focused business process management company, to help it determine how to rethink and redesign its roadmap to capture missed revenue opportunities.
WNS determined that the agency lacked an adequate offline strategy, which resulted in a dip in revenue and market share. After a deep dive into customer segments, the performance of offline sales agents, ideal hours for sales agents, and more, WNS was able to help the agency increase offline revenue by 16 percent and increase conversion rates by 21 percent.
5. How Mint.com Grew from Zero to 1 Million Users
Mint.com is a free personal finance management service that asks users to input their personal spending data to generate insights about where their money goes. When Noah Kagan joined Mint.com as its marketing director, his goal was to find 100,000 new members in just six months. He didn’t just meet that goal. He destroyed it, generating one million members. How did he do it?
Kagan says his success was two-fold. This first part was having a product he believed in. The second he attributes to “reverse engineering marketing.”
“The key focal point to this strategy is to work backward,” Kagan explained. “Instead of starting with an intimidating zero playing on your mind, start at the solution and map your plan back from there.”
He went on: “Think of it as a road trip. You start with a set destination in mind and then plan your route there. You don’t get in your car and start driving without in the hope that you magically end up where you wanted to be.”
6. Netflix: Using Big Data to Drive Big Engagement
One of the best ways to explain the benefits of data science to people who don’t quite grasp the industry is by using Netflix-focused examples. Yes, Netflix is the largest internet-television network in the world. But what most people don’t realize is that, at its core, Netflix is a customer-focused, data-driven business. Founded in 1997 as a mail-order DVD company, it now boasts more than 53 million members in approximately 50 countries.
If you watch The Fast and The Furious on Friday night, Netflix will likely serve up a Mark Wahlberg movie among your personalized recommendations for Saturday night. This is due to data science. But did you know that the company also uses its data insights to inform the way it buys, licenses, and creates new content? House of Cards and Orange is the New Black are two examples of how the company leveraged big data to understand its subscribers and cater to their needs. The company’s most-watched shows are generated from recommendations, which in turn foster consumer engagement and loyalty. This is why the company is constantly working on its recommendation engines.
The Netflix story is a perfect case study for those who require engaged audiences in order to survive.
In summary, data scientists are companies’ secret weapons when it comes to understanding customer behavior and levering it to drive conversion, loyalty, and profits. These six data science case studies show you how a variety of organizations—from a nature conservation group to a finance company to a media company—leveraged their big data to not only survive but to beat out the competition.