The Evolution of Data-Driven Decisions: Predictive Analytics

By: Katana Graph

March 19, 2022

The Evolution of Data-Driven Decisions: Predictive Analytics

The term “analytics” encompasses an entire realm of thought about processing data. There are several different kinds of analytics contained within the overarching term, but all analytics strive to gain some sort of understanding from data. Modern analytical methods tend to be lumped into four groups: descriptive, diagnostic, predictive, and prescriptive.

Descriptive analytics chronicle events that have taken place and we use diagnostic analytics to determine why those events have taken place. Predictive analytics consider all known information about a given subject to make predictions about future events. The value of predicting future outcomes and planning for unknown events is obvious, but reaching these outcomes is far from straightforward. Taking into account all information about a given subject while considering the volume and variable types of data available to process and produce these outcomes is overwhelming.

Predictive analytics dates back to the 1700s, as governments began keeping better records of births, deaths, and crimes. As patterns in the data emerged, predictions such as how many deaths would occur in Paris could be made, but inferring who would pass away and how they did so remained unpredictable. As time went on and data collection became more organized, it became apparent that identifying the reason behind individual events could be applied to recognizing patterns in larger groups. Trends in public health and criminal behavior emerged and were used to help serve societal interests. Predictive analytics in business describe the use of statistical data and modeling techniques to project future outcomes, risks, and economic performance.

Using an example from the realm of home economics, suppose we want to try a new cookie recipe this holiday season but we’re unsure whether it will be a hit. Should we choose to create our own recipe, we might need to understand how molecular bonds form between ingredients in addition to how they react to heat in the oven. For example, evidence shows that to produce a more consistent and repeatable cookie result, the baker should mix wet and dry ingredients separately to evenly disperse rising agents and flavor choices.

First, we could identify desirable ingredient choices and their repercussions. If we want to disappoint most children, we should use raisins instead of chocolate chips. If we want to accommodate a gluten-free friend, we’ll need something to stand in as a binding agent for the batter. Data analytics can tell us how and why the inclusion of walnuts versus carrot shavings influences the baking heat, time, and choice of icing in addition to the ideal volume of walnuts or carrot shavings to include.

These are fairly concrete examples; predictive analytics would also take into account things like an individual’s familiarity with a particular cookie. Perhaps for several decades, one family member has brought to every family event exceptionally dry and crumbly cookies with so many walnuts in them that the cookie was considered by family members a dessert item by name alone.

Predictive analytics takes into account population trends and the subject’s membership in the reference class of homesick family members, which in this instance tells us that if a family member who hasn’t been home in a few years eats one of these miserable cookies they will enjoy it despite it being an otherwise undesirable food item. Prescriptive analytics in contrast, as discussed in a future post, tells us that we should ship our distant family members some cookies.

Zooming out to the bigger picture of food preparation and distribution, predictive analytics uses customer buying patterns, regional preferences, and supply chain details to predict stocking needs, employee staffing schedules, and retail profit margins. Zooming further out, we see AI-based predictive analytics at work across the entire spectrum from farm to table. Farms provide soil test results, sometimes collected by IoT devices, along with planting and harvesting times into analytics tools where it is combined with weather data for the entire growth cycle of the crop. Produce packagers supply sorting, washing, packaging, and tracking data. Shippers contribute processing times and routing details, all of which enter into predictive analytics to forecast storage needs, shelf life, spoilage, and customer demand.

Get to Know Katana Graph

We thrive on testing new and diverse ideas.

Katana Graph was born of cutting-edge research and scientific rigor, and these beginnings have a powerful effect on who we are to this day. We’re devoted to problem solving, and are relentless in our pursuit of more effective and more efficient solutions to real-world challenges. Continuous improvement is the very foundation of our success.

Whether structured or unstructured data, Katana Graph can greatly improve the insights and opportunities within your organization’s data. Click to schedule a meeting.


Newsletter Sign Up

Subgraph Extraction

Graph mining is a growing field of study, and its use applies to many real-world applications. In.

Read More
ETL with the Katana Graph Library

ETL (extract, transform, and load) refers to the process data engineers use to pull data from.

Read More
What Is PageRank?

The PageRank algorithm uses incoming links between a graph’s nodes to rank the nodes, giving nodes.

Read More

View All Resources

Let’s Talk

Turn Your Unmanageable
Data Into Answers

Find out how Katana Graph can help provide the foundation for your future of data-driven innovation.

Contact Sales