Data Science

Projects and Customers...

Forecasting Churn and
customizing Loyalty Actions

AT&T

Keywords: Journey Analytics, Churn Analysis, Customer Lifetime Value, Dynamic Segmentation, Clustering, Pattern Matching

 

Calculating how much to spend on acquisition and retention is something of a black art. If too many customers are lost, revenues will plummet. If too much is spent, margins will suffer unduly. But big money is at stake.

Recent research studies show that the average service provider in a mature market typically spends 15-20% of service revenues on acquisition and retention activities.

Discover how Machine Learning can help...

RAN Anomaly Detection

SPRINT

Keywords: Anomaly Detection, Behavioural Analysis, Logistic Regression, Kafka, Spark, spark Streaming, Unsupervised Machine Learning, Fuzzy Clustering, Time Series Analysis, Arima, Entropy Index, Index of Dispersion, Correlation Index (Pearson, Kendall, Spearman), MLib.

The complexity of mobile Radio Access Networks is quickly increasing and demanding additional efforts to monitor and maintain. Thingbook learns network data’s behaviour to automatically detect anomalies, correlate Key Performance Indicators (KPIs) behaviours, discover new patterns and perform root-cause analysis helping Telcos to optimize service levels, minimize overhead costs and maximize profitability.

Classification of Device Behaviour in Internet of Things Infrastructures

TU DUBLIN

Keywords: Apache Kafka, Spark Streaming, Internet of Things, Cyber Security, Streaming Analytics, Device Behaviour Classification, Abnormal Behaviour Detection

 

We are fast approaching the point where the existential problem of IoT based cyber-security attacks is a serious threat to industrial operations, business activity and our social interactions that leverage IoT technologies...

Our Technology applied a Spatio-temporal methodology to characterise network behaviours. Once characterised, anomalous behaviours can be identified by calculating a similarity metric to previously identified behaviours (e.g. attacks, intrusions or malfunctioning machinery)

Unsupervised Fraud detection for Insurance

VHI

Keywords:  Insurance, Fraud detection, Deep Learning, LSTM, CNNs, Auto-Encoders, Pattern recognition, Claim processing

 

"The extent of insurance fraud varies between countries. Detected and undetected fraud is estimated to represent up to 10% of all claims expenditure in Europe. This figure varies between countries and classes of insurance due to a number of factors, such as how the market functions or the local prevalence of one type of insurance. " Insurance Europe

 

Our Technology stores, analyses and evaluates claims and customers behaviours simultaneously, finding anomalies and similarities and reducing the human intervention in problems like fraud detection, the discovery of new ways of fraud and accurate policy personalization.

Network Intrusion Detection

GLOBE

Keywords: Anomaly Detection, Behavioural Analysis, Kafka, Spark, spark Streaming, Unsupervised Machine Learning, Fuzzy Clustering, Time Series Analysis, Arima, Entropy Index, Index of Dispersion, Pearson correlation, MLib and R

 

The increasing adoption of connected devices produces a massive volume of streaming, time-series data presenting significant technical challenges for telecom operators in areas like Quality of Experience, determination, Resource Scalability, Predictive Maintenance, Fraud Detection and Cyber Security.

 

One fundamental capability for network streaming analytics is to model each stream in an unsupervised fashion and detect unusual, anomalous behaviours in real-time as training batch are not always available, particularly when new network behaviours emerge.

Services...

Get the extra help you need in your data science projects with seasoned data scientists. We accumulate a strong experience working on use cases requiring Anomaly Detection, Time Series Analysis and Prediction capabilities using Deep Learning, ML and Statistics Methods.

 

The core of our experience is based on the following areas: Regressive models (ARMA, ARIMA, SARIMA), LSTMs, Auto-Encoders, Boltzmann Machine,  Fuzzy/Density/Hierarchical  Clustering and Regressions using popular tools as R, Phyton, Keras, Tensor-flow, Scikit-learn and Numpy.

Data Science as a Service

Crafting and executing the data strategy for your organisation is  critically important to your business. Thingbook CDO-Smart Services will craft a business aligned data strategy which supports and enables your business ambitions, resolves pain points and unlocks latent data potential.

 

The strategy is presented and aligned with your stakeholders to bring all constituents into agreement. Our service creates business cases for a data strategy to garner investment and priority from leadership/boards. With an appropriate data strategy in place, we will help you to plan, execute and manage the implementation and drive adoption.

Data Strategy Consultancy

© ThingBook.io - All rights reserved

  • White Twitter Icon
  • White LinkedIn Icon