Ready-made solutions bases on
Big Data / Machine learning / Data Science
Our solutions in industry context
Solutions for Retail
Product range optimization
Personal product offerings
Logistics optimization, purchace timing and goods warehousing in stores
Logistics optimization, purchace timing and goods warehousing in stores
Market indexes forecasting for the new products
Working hours optimization
Loyalty programs optimization
Brand perception monitoring in mass media and internet
Retail network locations optimization
Optimizataion and automatic control of products merchendising
Customers classification and / or customers' behaviour classification based on in-store video records
Working hours optimization
Call center optimization
Voice menue
Typical questions robotic response
Solutions for the Banks
Customer Live Time Value estimation
Customers segmentation
Product recommendations for the customers / cross-sale
Demand forecasting in the contexts of bank products or customers
Next best offer for the customers
Credit default risk prediction (B2B/B2C)
ATMs cash demand forecasting / free capital demand / bank reserves
Affiliated customers revealing
ATM machines/bank offices locations optimization
Churn aptitude prediction
Pricing strategy analysis (for example when price decrease may lead to significant demand increase and boost profits)
Fraud transactions tracking for automatic credit card blocking
Call center optimization
Operators work hours optimization bases on workload prediction
Operators work hours optimization bases on workload prediction
Solutions for Transport
Malfunctions prediction for preventive maintenance
Real time manufacturing proceses optimization - where and how many resources to allocate
Routes and timetables optimization
Optimal type of transport selection for the corresponding route
Outstanding events detection (non-typical working operations, rout deviation, etc.)
Detection of typical reasons that lead to quality standards violations (malfunctions, deleays, etc.)
Market indexes forecasting for the new products/services
Personal product/services offerings
Market demand forecasting (shipment volumes) and procurement planning
Dinamic pricing
Call center optimization
Solutions for Manufacturing
Expensive raw materials volume optimization (additives, reagents) to decrease production costs without the quality changes
Production parameters real time optimization. For example decrease of electricity consumption without production volume decrease
Prediction of proceses parameters, which are impossible or hard and expensive to measure based on known factors
Prediction of expected goods quality and possible defects early in the early production stages. Useful for decrease of overall defects and optimization of materials processing routes
Why does it make sence to work with us?
We solve your business problems
We work with both:
big businesses and small teams
big businesses and small teams
We implement up-do-date
workflow methodologies
workflow methodologies
Our pricing
is highly competitive
is highly competitive
Our team has a proven
highly qualified expertise
highly qualified expertise
Our executed projects results
75% average decrease in warehousing stock
Objective:
Decrease of items stock in stores avoiding the "out of stock" situation on the shelves
97% accuracy of sales volume interval prediction VS actual sales
Solution:
The developed model predicts weekly sales volume of each item for each week of the year
Tools: R, Python
Algorithms: XGBoost, CART, k-means
10% maximum diviation
from expert forecast
Objective:
Head office financial planning automation by means of regional budgetary expences forcasting
over 120 cases
of fraud revealed
Solution:
A custom solution based on KXEN and SAP BI had been developed and implemented within customer's infrastructure. As a result, the regional subsidiaries' expences are being forecasted based on their historical costs statistics and current economic envorinment. In additional fraud detection mechanism is implemented which inputs these unexpected costs into head office's financial plan
Tools: R, KXEN, SAP BI
Algorithms: KXEN time-series algorithms
9 is the minimum quantity of clusters with clearly defined borders for targeted marketing approaches
Objective:
Differentiation of customers into groups for targeted marketing. Detection of customers who also use other providers' services
86% accuracy in customer classification, who also use competitive services
Solution:
Development of a custom analytical tool that divides the customer base into clusters defining each group's features to support the marketing department decision making. Separation of customers who also use competitive services
Tools: R, Python, Spark, Hadoop, Hive
Algorithms: XGBoost, CART, k-means, k-medoids, t-SNE
8 different data sources
Objective:
Tracking and prevention of oil refinery products losses (in volume and quality)
revealed over 1000 instancies
Solution:
Big Data platform based on Hadoop stack had beed deployed from scratch and integrated with data sources. Developed and implemented ETL (Extract, Transform, Load) processes. Created tracking and analytics modules which, together with other tools are available via the created web application
Tools: Python, JavaScript, Shell, SQL, Hadoop, HDFS, Spark, Yarn, Hive, Tez, Hue, Zookeeper, Luigi, Kafka, NiFi, PostgreSQL, Microsoft SQL Server, Oracle, Pi System, Django, AngularJS, Bootstrap, CSS, Docker, CentOS/RHEL
over 80% accuracy of the actual disease to fall in the TOP 3 predicted deceases for a patient
Objective:
Development of a tool to assist doctors diagnose patients based on data that is automatically collected from the devices monitoring the patient
98% probability of a patient not to have the disease with minimum predicted probability
Solution:
The developed model automaticaly predicts the diseases probabilities form a given group for a patient based on data from health monitors
Tools: R
Algorithms: GBM, AdaBoost, k-means
Development in three stages
1 Preparatory stage
Data assessment
Determination of project requirements, metrics and success criteria
Authorisation of experiment methodologies
Data transfer
2 Pilot results verification
Model creation
Model testing (experiment execution)
Success criteria matching, model effectiveness testing
3 Utilization
Coordination of regular data transfer and model deployment within the customer infrastructure
Timeley A/B testing and service quality control
Support and maintenance including the model update
We use only the best technologies
