Predictive maintenance

Anticipate on important events

Preventing problems instead of solving problems enables an organization to save a lot of money. With this in mind, BigData Republic has aided the development of predictive maintenance capabilities regarding asset management for a large electricity- and gas distribution network operator in the Netherlands. Using machine-learning techniques, we predict failure probabilities of individual network assets based on their characteristics.

The predictions from these data-driven models support domain expert decision-making on the short term, and feed into resource-allocation optimization models on the long term. By deploying predictive maintenance routines, the network operator is able to reduce both repair costs and failure risk.


An offer you can’t refuse

Every customer is unique. Instead of relying on business rules or coarse segmentation, we advocate the use of big data to make personalized offers to customers. BigData Republic has developed several big data recommender engines for retail companies. We have developed machine-learning models that can profile customers and predict the preference of a customer for a certain product based on multiple data sources.

The actionable insights provided by the engine are employed to increase conversion rate, sales and profits, by making more relevant recommendations to their customers.


Security analytics

For an oil major we built a security analytics platform storing over 1 Petabyte of data. Security analysts are now able to quickly search the data and will get real-time alerts based on common anomaly patterns detected in real-time. Major challenges include making the vast amount of data searchable and storing it securely in a public cloud.

Our solution has significantly reduced the time it takes before security incidents are discovered, and it enables for fast proof collection in audit trails.

Predicting Passenger Behavior

for large airline companies, information about its passenger is key to its operation decisions. With this in mind, BigData Republic has worked on several use-cases to improve operation efficiency, increase profit and decrease costs. Examples include: implementing a chatbot for customer services, delay prediction, forecast the number of no-shows and the prediction of no-shows, the amount of luggage, auxiliary sales and the amount & type of meals on-board.

A perfect example of a variety of experiments for one company that proved the value of data science, to be exploited in the next phase.


Real-time alerts for suspicious transaction

For a major bank we built a real-time fraud detection system that can signal suspicious patterns in bank transactions. Our natural language processing models are able to detect when a transaction contains a person or company who is sanctioned and is not allowed to receive any money. The solution will group anomalies of similar nature and present them to the analysts to decide whether or not to deny these transactions.


Tailored use cases to get the most out of your data

The use-cases outlined here are merely examples of the business value that big data analytics can add to an organization. Whether it be data from mobile applications, car on-board sensors, job market evolution, or data regarding energy production and distribution – BigData Republic offers the perfect kick-start for your big data potential.