Have you ever wondered why most of the ads popping up on your social media are relevant to you or why most films that your favorite streaming service suggests are actually your kind of movie? Personalized online advertising and movie suggestions are both produced by the same data science technology. Hadley Wickam, Chief Scientist at the data tools company RStudio, likens the process to gaining insight from raw data to derive useful and practical information relevant to the target audience. For advertisers, data science provides actionable intelligence on their customers.

Browsing the web, users leave behind traces of data such as location, age, search history, and more. Companies that place ads collect these digital footprints, cross-reference them to each other and then personalize the type of ads best suited to the individual web user. “The cool part of this technology is that the whole process is automated,” says Home Credit Indonesia Head of Data Science, Kirill Odintsov. “Data science isn’t about knowing methods or algorithms. It’s about forming a hypothesis, finding data to test it and reaching a good enough understanding of the problem to make actionable decisions with impact,” says Kirill.

For social media users or subscribers of streaming services, relevant ads or suggestions are just a convenience to check the offers that might be relevant to their interests. To sellers though, they provide a very important tool in understanding what their existing or potential customers want or need.

Data science is continually improving

‍Data science is not new. In the early days, data scientists were simply called statisticians or analysts. The underpinnings of most methods used by data scientists today aren’t as new as one would think. Simply the current achievable computing power is advanced so much, allowing to start using these methods on a massive scale. To be sure, some specific methods such as Extreme Gradient Boosting are relatively young, not even 10 years old. The real key difference is that modern-day data scientists can build their models much more quickly than their predecessors thanks to ever more sophisticated computing technology, including continually improving super computers.

Kirill says that Home Credit also utilizes data science to help the company achieve its various targets. As a financing company, Home Credit deals with a lot of transactions that make it imperative for the company to understand patterns in order to prevent unnecessary losses. “Home Credit also applies data science to help the company achieve its goals,” says Kirill. As a consumer credit company, Home Credit handles a lot of transactions that make it vital to understand the many different patterns in these transactions to prevent loss. For starters, Home Credit uses data science to detect frauds, such as false transactions, which can be made look like purchases to boost a users’ credit worthiness.

“Our anti-fraud data monitoring is basically about trying to prevent fraudulent activities designed to steal money from the client,” Kirill says. “Machine learning tracks these patterns to identify suspect transactions and then notifies us automatically. It is doing so while learning by doing to improve its detection prowess.”

Data science models can be programmed to understand the information on credit cards, and cross-reference them with data of the card owner data, locations, or time. By sifting through this complex mix, Home Credit can identify strange transaction to identify potential fraud swiftly and decide on the correct action to take.

Data scientists meet customer needs

‍Home Credit’s customer relation management system, or CRM, also functions best with data science support. As a system designed to serve customers, the CRM provides relevant information to customers based on parameters such as age, gender and location.

Data scientists can follow customer needs and offer them the most relevant product. But beyond placing ads, data can also give insight into customer behavior. Researchers can learn when the best time to call customers with personalized offers is or find out which communication platform is best to use to make an offer. “Relying on the understanding derived from data to gauge customer backgrounds and credit ratings, the Home Credit risk team is a key user of data science applications,” says Kirill.

Since nearly all businesses currently use some form of data science, it is no surprise that demand for data scientists continues to grow and outstrips their supply.

If computer modeling and data processing are your passion, Kirill has a few suggestions for you to start building your career as a data scientist. First, refresh your knowledge of statistics by going over your notes from college or other lessons. If you don’t work in this field, chances are your statistics might be a bit rusty or gone altogether. You need to sharpen your statistics skills. Learn Python or R programming languages. They are the two basic coding languages crucial to starting your journey in the world of data science. Also don’t be shy to take part in coding competitions such as run by Kaggle or others. The best way to learn data science is by doing it, same as the case is for any other skill.

“These are my quick tips if data science is really your thing and you want to become a true data scientist,” he says, adding that mastering the two key skills – statistics and programming languages – will give you the best chance to help companies maximize their business potential. Curiosity and being able to explain how you collected the data, analyzed the finding and predicting what will happen in the future are also highly effective skills.